up,

,n.wv«.<q...-o

gnu-neu-

w: w-

,» ~f “--

 

 

 

 

.-_
w“

 

“"3-
..

 

 

 

 

 

 

 

 

 

 
 

 

 

 

  

 

 

 

 

 
 

 
 

 

 

 
 

 

 

Z LIBRARY
.2190? Michigan State
niversity

This is to certify that the
dissertation entitled

THE ROLE OF NON-CONSUMPTIVE EFFECTS IN THE NET
EFFECT OF AN INVASIVE PREDATOR IN THE
LAURENTIAN GREAT LAKES

presented by

KEVIN L. PANGLE

has been accepted towards fulfillment
of the requirements for the

Ph.D. degree in Fisheries & Wildlife
Ecology, Evolutionary Biology,
8. Behavior

 

 

//U’Zf ,e—Jd

Major Prdfessor’s Signature
5’ ' /5‘- 0?

Date

 

MSU is an afiinnative-action, equal-opportunity employer

.-.—.-.-.--o-.--n-u—I-nu-.-.—--o--.--.-o-.--.----o-.------—.--.

no-o-n-u—-----o--.--.--.---a--:-—-.—t

 

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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K:IProlecc&Pres/ClRC/DateDue indd

THE ROLE OF NON-CONSUMPTIVE EFFECTS IN THE NET EFFECT OF AN
INVASIVE PREDATOR IN THE LAURENTIAN GREAT LAKES

By

Kevin L. Pangle

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Fisheries and Wildlife
Ecology, Evolutionary Biology, and Behavior

2008

ABSTRACT

THE ROLE OF NON-CONSUMPTIVE EFFECTS IN THE NET EFFECT OF AN
IN VASIVE PREDATOR ON LAURENTIAN GREAT LAKES FOOD WEBS

By

Kevin L. Pangle

In ecological food webs, predators can have consumptive effects (CBS) and non-
consumptive effects (N CEs) on prey populations, the latter caused by inducing costly
changes in prey phenotypes. Studies of NCEs, primarily carried out in laboratory and
mesocosm settings, have shown that NCEs can profoundly influence the structure and
dynamics of food webs. Whether or not NCEs contribute significantly to ecological
patterns and processes observed in the field is still an open question, but one essential to
address in the aim of better understanding the mechanisms that regulate food web
dynamics in nature and to better inform the management of ecosystems. My goal in this
dissertation was to elucidate the importance of NCEs in the complex, natural setting of
the Laurentian Great Lakes. I focused my research on the NCES of the invasive predator,
Bythotrephes longimanus, and in doing so, I hoped that my dissertation work would also
aid in better understanding the detrimental impact of this invader.

The approach I took was necessarily very different than those taken by previous
studies of NCEs. Unlike laboratory and mesocosm settings, the Laurentian Great Lakes
are not conducive to experimental manipulations. Instead, to evaluate NCEs, I relied on a
sequence of steps that integrated experimental findings, field observations, and modeling.
First, I tested the phenotypic response of the native prey (zooplankton) to Bythotrephes

and related their response to zooplankton phenotypes observed in the field. Second, I

evaluated the possible consequences of the phenotypic response of zooplankton on their
population growth rate (i.e., NCEs of Bythotrephes) in the field. Third, I incorporate
NCEs into food web models to quantify the contribution of NCEs to the net effect of
Bythotrephes on zooplankton populations and to the net indirect effect of Bythotrephes on
an ecologically and economically important Great Lakes species, alewife (Alosa
pseudoharengus).

I found that multiple Great Lakes zooplankton species respond to Bythotrephes by
modifying a behavioral trait. Specifically, the presence of Bythotrephes induced
zooplankton to migrate downward, which I observed in the field and tested in laboratory
settings. Vertical migration is assumed to be adaptive as it reduces the spatial overlap
between zooplankton and Bythotrephes. To avoid Bythotrephes, I found that zooplankton
incur a substantial cost to their population growth rate by shifting habitat use to cooler
environments. Estimates of the magnitude of these NCEs in Lakes Michigan and Erie
indicated that NCEs play a large role in the effect Bythotrephes has on zooplankton
populations, with NCEs on the same order of magnitude as or greater (up to 10 fold) than
CEs. Further, consideration of NCEs greatly influenced the predicted indirect effects of
Bythotrephes on alewife growth and survival in Lake Michigan.

These results demonstrate the relevance of NCEs to ecological patterns and
processes observed in nature. Future consideration of NCEs may substantially improve
our understanding of food web function and better inform the management of ecosystems
that rely on this understanding. These results also demonstrate the complex impact the

introduction of species can have on ecosystems through NCEs.

Copyright by
KEVIN L. PANGLE

2008

ACKNOWLEDGEMENTS

Many people have been instrumental to the completion of this dissertation. I
would like to first thank Dr. Scott Peacor, who has played the role of teacher, mentor,
collaborator, and friend and who has broadened my horizons to see the beauty of basic
science. He, along with my other committee members, James Bence, Andrew McAdam,
and Orlando Samelle, have kept me on the right path during my time at Michigan State
University by graciously lending their time and expertise. For all the field and lab work,
I relied on help from undergraduate assistants, and I feel fortunate to have interacted with
such a great bunch of students. They include Kurt Anderson, Christine Bergeon, Jason
Fischer, Katie Krieger, Dianna Miller, Jason Williams, and Angela Woods. I also
received a tremendous amount of assistance from personnel at the National Oceanic and
Atmospheric Administration (N OAA) Great Lakes Environmental Research Laboratory.
I would like to particularly acknowledge Dennis Donahue and the NOAA ship crew,
including Bill Burns, Jack Workman, and Andrew Yagiela, who were extraordinary
generous with their time and effort.

Being a part of several scientific communities while at Michigan State University
(MSU) has been a priceless asset to me, both in the development of the research in this
dissertation and, more generally, in my growth as a scientist. First, I am grateful for the
interactions I have had with the Bythotrephes research community, particularly fi'uitful
discussions with Ora Johannsson, Norman Yan, and Joelle Young. Second, I feel very
lucky to have interacted with my lab mates in the Limnology Lab at Michigan State

University (MSU), who I have mercilessly bombarded with thoughts and ideas and who

were always willing to lend a helping hand. They include Tom Alwin, Josh Booker, Katie
Button, Kendra Cheruvelil, Emi Fergus, Geoff Horst, Lesley Knoll, Drew Kramer, Sherry
Martin, Andrea Miehls, and Emily Norton. Finally, I am thankful for working in the
positive and constructive environment fostered by the Department of Fisheries and
Wildlife and the Ecology, Evolutionary Biology, and Behavior Program.

Through this endeavor, I have remained fueled by the love and support of family
and friends. I would like to particularly acknowledge my parents, Chris and Wendy
Pangle, for the high value they have placed on education, their encouragement to pursue
dreams, and their belief in me. Last, but not least, I would like to thank my wife, Wiline
Pangle. She has been a true partner in all of this and a constant source of inspiration.

Funding for the research in this dissertation was provided by the Fishery Research
Program of the Great Lakes Fisheries Commission, NOAA, Environment Canada, the
Michigan Agricultural Experiment Station, the Cooperative Institute for Limnology and
Ecosystems Research, the MSU Department of Fisheries and Wildlife and the MSU

Ecology, Evolutionary Biology, and Behavior Program.

vi

TABLE OF CONTENTS

LIST OF TABLES .................................................................................. ix
LIST OF FIGURES ................................................................................. x
GENERAL INTRODUCTION .................................................................... 1
CHAPTER 1
NON-CONSUMPTIVE EFFECT OF THE INVASIVE PREDATOR BY T HOT REPHES
LONGIMAN US ON DAPHNIA MENDOTAE .................................................. 10
Abstract ..................................................................................... 10
Introduction ................................................................................. 10
Methods ..................................................................................... 13
Results ....................................................................................... 1 8
Discussion .................................................................................. 22
CHAPTER 2 '
LACK OF NAIVETE IN LAKE MICHIGAN ZOOPLANKTON PREY TO RISK
POSED BY INVASIVE PREDATORS . . . . ........................................................... 27
Abstract ...................................................................................... 27
Introduction ................................................................................. 28
Methods .................................................................................... 30
Results ....................................................................................... 39
Discussion .................................................................................. 44
CHAPTER 3

TEMPERATURE GRADIENTS, BUT NOT FOOD RESOURCE GRADIENTS,
AFFECT GROWTH RATE OF DAPI-HVIA MENDOTAE IN LAKE MICHIGAN.........48

Abstract ...................................................................................... 48
Introduction .................................................................................. 49
Methods ..................................................................................... 51
Results ....................................................................................... 58
Discussion ................................................................................... 60

CHAPTER 4

LIGHT-DEPENDENT PREDATION BY THE INVERTEBRATE PLANKTIVORE,

BYT HOT REPHES LONGIMAN US ............................................................... 65
Abstract ...................................................................................... 65
Introduction ................................................................................. 65
Methods ..................................................................................... 68
Results ....................................................................................... 75
Discussion ................................................................................... 77

vii

CHAPTER 5
LARGE NON-CONSUMPTIVE EFFECTS OF AN INVASIVE PREDATOR ON

ZOOPLANKTON POPULATION GROWTH RATE ......................................... 85
Abstract ...................................................................................... 85
Introduction .................................................................................. 86
Methods ..................................................................................... 88
Results ....................................................................................... 96
Discussion ................................................................................. 102

CHAPTER 6

INTERACTIONS BETWEEN LAKE MICHIGAN COMPETITORS MEDIATED BY

THE BEHAVIOR OF A SHARED PREY ..................................................... 109
Abstract .................................................................................... 109
Introduction ................................................................................ 1 1 0
Methods .................................................................................... 113
Results ...................................................................................... 126
Discussion ................................................................................. 132

APPENDIX A ........................................................................................... 135

APPENDIX B ........................................................................................... 136

APPENDIX C ........................................................................................... 138

APPENDIXE ........................................................................................... 141

LITERATURE CITED ............................................................................. 143

viii

LIST OF TABLES
Table 2.1. Names and characteristics of the native and invasive plantivores used in the
experiments ........................................................................................................................ 32

Table 3.1. Temperature and food resource treatments used in the first and second
experiments. ...................................................................................................................... 56

Table 4.1. Summary of the different Bythotrephes densities and light intensities used for
each treatment in Experiments 1 and 2 .............................................................................. 69

Table 4.2. Bythotrephes clearance rates (L hr") and ingestion rates (number hr'l) of
Daphnia mendotae observed in Experiments 1 and 2 at different intensities (description
of light intensities in Table 1) ............................................................................................ 78

Table 5.1. Sampling locations and dates for Lake Michigan and Lake Erie field
surveys ............................................................................................................................... 92

Table 6.1. Eight model scenarios that included (X) and excluded (0) different indirect
effects of Bythotrephes on alewife ................................................................................... 116

Table 6.2. Summary of characteristics of each prey type used in the model ................... 121

Table 6.3. Results of sensitivity analysis, in which the overwinter survival in the "full
effect" (original 18 %) was estimated after a i 50 % in each parameter ......................... 131

ix

LIST OF FIGURES

Figure 1.1. Results of Experiment 1. Day (a) and night (b) vertical position, plotted as a
function of depth (left hand axis) and temperature (right hand axis), of D. mendotae
exposed to aged lake water without (solid circles [0]) and with (open circles[O])
Bythotrephes kairomone. The left-hand frequency distributions represent the mean
percentage (3: SE bars) of Daphnia at each depth interval (points have been positioned on
the y-axis at the midpoint of their respective depth interval). Mean percentages for the top
2 5-cm intervals have been combined, as have those for the bottom 2 5-cm intervals. The
right-hand points represent the mean depth (i SE bars) .................................................... 19

Figure 1.2. Mean depth (a) and the percentage of D. mendotae in the upper and lower 5-
cm extremes (b and c, respectively) in Experiment 2. Treatments included: control water;
lake water not inhabited by Bythotrephes; lake water inhabited by Bythotrephes
(approximately 0.010 Bythotrephes L'l); low Bythotrephes-density laboratory-prepared
water (0.05 Bythotre hes L'l); medium Bythotrephes-density laboratory-prepared water
(0.5 Bythotrephes L' ); and high Bythotrephes-density laboratory-prepared water (five
Bythotrephes L"). Values reported are treatment means (:t SE). Means with different
lowercase letters were significantly different as determined by ANOVA and Tukey’s
HSD tests (p < 0.05) .......................................................................................................... 21

Figure 1.3. Observed (open bars) and predicted (using a growth model, hatched bars)
specific growth rate of D. mendotae in Experiment 1 exposed to aged lake water without
and with Bythotrephes kairomone. Values reported are treatment means (i SE) ............. 25

Figure 2.1. Mean depth (panel a) and the percentage of D. mendotae in the upper and
lower 5-cm extremes (panels b and c, respectively) in Experiment 1. Values reported are
treatment means (:1: SE). Treatments included: control water or water with kairomones
from the invasive predators Cercopagis (CP) or Bythotrephes (BL) or from the native
predators, Polyphemus (PP), Leptodora (LK), or Perca (PF). Asterisks indicate predator
treatments that were significantly different (p < 0.05) from the control ........................... 40

Figure 2.2. Mean depth (panel a) and the percentage of D. mendotae in the upper and
lower 5-cm extremes (panels b and c, respectively) in Experiment 2. Values reported are
treatment means (3: SE). Treatments included: control water or water with kairomones
from the native, deep-dwelling predator, Mysis, or the invasive, shallow-dwelling
predator, Bythotrephes. Asterisks indicate treatments that were significantly different (p <
0.05) from the control ........................................................................................................ 42

Figure 2.3. Mean depth (panel a, b) and the percentage of individuals in the upper and
lower 5-cm extremes (panels c,d and e,f respectively) in Experiment 3. Left panels give
results for Bosmina longirostris and right panels give results for Daphnia pulicaria.
Values reported are treatment means (:1: SE). Treatments included: control water or water
with kairomones from the small invasive predator, Cercopagis (CP), or the large, native

predator, Perca (PF). Asterisks indicate treatments that were significantly different (p <
0.05) from the control ........................................................................................................ 43

Figure 3.1. Water temperature (solid line) and in situ chlorophyll fluorescence (dashed
line) in Lake Michigan during water collection for the first and second experiments
(panels a and b, respectively). Arrows indicate the depths at which food resources were
collected from for each experiment ................................................................................... 53

Figure 3.2. Growth rates of Daphnia in Experiment 1. Juvenile Daphnia experienced
three temperature treatments (8°C, alternating 8 and 21°C, and 21°C) and two food
resource treatments (water from the epilimnion and the deep chlorophyll maxima
[DCM]). Values reported are treatment means (d: SE) ...................................................... 59

Figure 3.3. Growth rates of juvenile Daphnia in Experiment 2. Daphnia experienced one
of four treatments including: deep chlorophyll maxima (DCM) water at 8°C for 12 h and
epilirnnetic water at 21°C for 12 h; epilirnnetic water at 21°C; epilirnnetic water at 21°C
for 12 h and at 8°C for 12 h; and DCM water at 21°C for 12 h and at 8°C for 12 h. Values
reported are treatment means (:I: SE). Means with different lowercase letters were
significantly different as determined by ANOVA and Tukey’s HSD tests (p < 0.05) ...... 61

Figure 4.1. Densities of different prey types in Bythotrephes and control treatments in
Experiments 1 and 2. In each experiment, bottles were exposed to different light
intensities and received 0, 1, or 3 Bythotrephes depending on the treatment (see figure
legends). Bars are means with standard errors and reflect total, not per capita predator
effect. Data from initial prey density are included in mean densities of the control
treatment. Prey types included Daphnia mendotae (D.m.), Bosmina Iongirostris (B.l.),
copepod nauplii, immature Diaptomus (I. Diap.), adult Diaptomus (A. Diap.), immature
Diacyclops thomasi (I. Diac.), and adult D. thomasi (A. Diac.). Asterisks denote
significant differences ....................................................................................................... 76

Figure 4.2. Effect of light intensity on Bythotrephes ingestion rate. Data of ingestion rates
of Daphnia mendotae in Experiments 1 and 2 were combined to evaluate Bythotrephes
predation over a broad range of light intensities. Points are means with propagated
standard error bars that were calculated using equations 1 and 2. In this analysis, the
relationship between light intensity and ingestion rate was described using a logistic
regression model (y = 2.12 / [1 + (x / 6.79)"'57]), shown by the fitted line ...................... 79

Figure 5.1. Percentage of B. longirostris (a, c) and D. mendotae (b, d) in the epi- and
hypolimnion of Lake Michigan over a gradient of Bythotrephes biomass (mg/m2). Lines
are linear regressions performed on transformed data and plotted on untransformed data,
with analysis results reported within each subplot. Stars (1%) represent the distribution in
the intensive survey (see text) ........................................................................................... 97

xi

Figure 5.2. Percentage of B. longirostris (a, c) and D. retrocurva (b, d) in the epi- and
hypolimnion of Lake Erie over a gradient of Bythotrephes biomass (mg/m2). Lines are
linear regressions performed on transformed data and plotted on untransformed data, with
analysis results reported within each subplot. Stars (731') and diamonds (0) represent
distributions in the first and second intensive surveys, respectively (see text) ................. 98

Figure 5.3. Vertical distribution of D. mendotae (solid line) and Bythotrephes (dashed
line), and mid-day temperature and light conditions of the water column (e) during the
Lake Michigan intensive field survey. Each of four sampling profiles are reported in
chronological order .......................................................................................................... 100

Figure 5.4. Vertical distribution of B. longirostris (solid line) and Bythotrephes (dashed
line), and mid-day temperature and light conditions of the water column during the first
and second Lake Erie intensive field surveys (a—c and d—e, respectively). Sampling
profiles were conducted near midday (a,d) and midnight (b,e) during both surveys ...... 101

Figure 5.5. Estimated lethal and non-consumptive effects of Bythotrephes on D.
mendotae in Lake Michigan (a) and B. longirostris in Lake Erie (b, c). Non-consumptive
effects (dotted line with open triangles) were estimated based on three different scenarios
(see Methods section) that assumed different prey distribution in Bythotrephes ’ absence
to reflect the natural variation. This range is represented on the x-axis as the percentage
of the prey species in the epilimnion in Bythotrephes ’ absence for the different

scenarios ........................................................................................................................... 103

Figure 6.1. Schematic of the two models used in this study and how they are linked to
each other ......................................................................................................................... 114

Figure 6.2. Seasonal environmental variables simulated in the model, including water
temperature (panel a), day length (panel b), and Bythotrephes density (panel c) ............ 115

Figure 6.3. Seasonal total density of Daphnia (panel a) and density of Daphnia available
for alewife to consume (panels b, c) under the eight model scenarios. Total Daphnia
density is not affected by reduced overlap with alewife, thus the eight model scenarios
are represented by four lines ............................................................................................ 127

Figure 6.4. Estimated mean growth rate and survival rate for alewife in the eight model
scenarios ........................................................................................................................... 129

Figure 6.5. Lengths of alewife over the season that were hatched on the first (solid line),
mean (dotted line), and last (dashed line) hatch date under the eight model scenarios...130

xii

GENERAL INTRODUCTION

The concept of food webs is of central importance to the field of ecology, as it has
been the foundation for understanding how interactions among organisms shape the
structure and dynamics of ecological systems (Elton 1927; MacArthur 1955; Paine 1980).
Ecologists have relied on models of food webs to elucidate processes that control
ecosystem stability, productivity, biodiversity (e.g., Rosenzweig 1971; May 1973;
McCann 2000), and to inform the management of ecosystem services (e.g., Garcia and
Cochrane 2003). To construct these food web models, the conventional approach relied
on linking predators and prey using pairwise interactions. These interactions are generally
defined as the per capita rate of consumption of the prey by the predator, referred to as
the predator’s consumptive effect (CE). Food web models therefore typically ignore
interactions among species in a food web that occur through means other than direct
consumption.

One ignored interaction is the effect predators can have on prey phenotype. As a
consequence of predation, many prey have evolved plastic anti-predation traits, including
behavioral, morphological, and life-history traits, that are modified by prey in a way that
optimizes prey fitness (reviewed in Lima 1998; Tollrian and Harvell 1999). As is general
in all forms of phenotypic plasticity, the expression of these anti-predation traits
represents a trade-off, in this case between the benefit of reduced predation risk and the
costs associated with the production and maintenance of the plastic trait. Prey are
predicted to incur these cost as long as the net benefit is maximal. Subsequently, these
costs, termed the non-consmnptive effect (N CE) of the predator on the prey (Abrams

2007), can be substantial and lead to significant reductions in prey somatic grth rate

and fecundity (Werner et a1. 1983; Peckarsky et a1. 1993; Diehl and Eklov 1995;
reviewed in Peacor and Werner 2004).

Trait modifications and consequent NCEs may influence how food webs function
in several ways. First, recent studies using mesocosm experiments in a damsel bug-pea
aphid system (Nelson et al. 2004) and in a Chaoborus-Daphm'a system (Boeing et a1.
2005) have found that the costs incurred due to trait modifications can lead to a
substantial NCE on prey population grth rate. This effect is qualitatively different than
a predator's CE in that prey population size is being reduced without directly increasing
predator population size. Consequently, NCEs may stabilize predator-prey dynamics
(Ives and Dobson 1987; Fryxell and Lundberg 1997), or destabilize predator-prey
dynamics if there is a time lag between prey and predator population changes (Luttbeg
and Schmitz 2000). Second, NCEs may introduce nonlinearities into predator-prey
functional relationships that are predicted to profoundly affect food web dynamics
(Abrams 1995; Bolker et a1. 2003). In particular, a small number of predators may have a
disproportionately larger effect on prey density if prey respond to their presence (Abrams
1995; Peacor and Werner 2004). Third, trait modifications that cause NCEs may
influence the interactions between prey with other species in the food web. This effect of
the predator is termed a trait-mediated indirect effect (Abrams 2007), which is a subset of
higher-order interactions (sensu Vandermeer 1969). Trait-mediated indirect interactions
can lead to counter-intuitive outcomes in food webs (reviewed in Werner and Peacor
2003), such as increased prey growth caused by increased predator density (Peacor 2002).

Although these studies have clearly shown the potential ways that NCEs may

influence food webs, whether or not NCEs are relevant to the ecological patterns and

processes observed in nature is still an open question. To my knowledge, when I started
my dissertation research, no study had quantified NCEs in an uncontrolled, field setting.
One obstacle facing such an endeavor is that the inherent complexity of ecosystems
makes measuring phenotypic response and their ensuing cost very difficult (Creel and
Christianson 2008). Another obstacle is that NCEs may have far reaching effects on the
possible multitude of other species in an ecosystem (Abrams 1995). Although
challenging, quantifying NCEs in the field may not only help identify mechanisms that
regulate food web dynamics in nature, but also inform ecosystem conservation and
management (Browman et al. 2004; Pikitch et al. 2004), where consideration for NCEs is
largely lacking (Creel and Christianson 2008).

My goal in this dissertation was to elucidate the importance of NCEs in the
complex, natural setting of the Laurentian Great Lakes. The Laurentian Great Lakes
provide a valuable natural resource for millions of people, supporting a large number of
commercial and recreationally activities. However, a number of serious problems
threaten this unique system. For example, invasive species are changing the structure and
functionality of native Great, Lakes communities (Vanderploeg et al. 2002), toxic algae
are closing beaches and polluting drinking water (Murphy et al. 2003), and important
fisheries are exhibiting drastic declines (Mills et al. 2003; Marsden and Robillard 2004).
A commonality among many of the diverse issues is their association with complex,
disruptive effects on the food web. Knowledge of food web interactions in the Great
Lakes may therefore be vital to gaining the most complete understanding of harmful

implications to the aquatic ecosystem.

An invasive species of particular concern to Great Lakes biologists is the
invertebrate species Bythotrephes longimanus. Bythotrephes is a predatory cladoceran
that feeds primarily on zooplankton (V anderploeg et al. 1993; Schulz and Yurista 1999).
Originally introduced via ballast water to Lake Ontario over 20 years ago, Bythotrephes
have spread to all five Great Lakes (Vanderploeg et al. 2002) and many surrounding
inland lakes (Boudreau and Yan 2003). Once established in a lake, Bythotrephes has been
credited with the local extirpation and reduction in density of several zooplankton species
(Lehman and Caceres 1993; Boudreau and Yan 2003; Barbiero et al. 2008). For example,
in Lake Michigan, populations of Daphnia mendotae, D. retrocurva, and D. pulicaria
collapsed during the first year Bythotrephes became abundant in the lake (Lehman 1991).
Daphnia mendotae have since rebounded to abundances that are equivalent to levels prior
to Bythotrephes invasion, while the other daphnid species have remained rare (Lehman
and Céceres 1993; Barbiero and Tuchman 2004). Such loss of zooplankton may in turn
have detrimental effects on many fish species for which zooplankton serve as a critical
food resource (Miller et al. 1990; O'Gorman et a1. 1997; Fulford et a1. 2006). This threat
spurred a recommendation to the Great Lake Fisheries Commission to determine the
effects of Bythotrephes on the pelagic food web (Shuter and Mason 2001).

Evidence suggests that Bythotrephes not only directly consume zooplankton, but
may also induce changes in zooplankton traits. Field studies indicate that the vertical
distribution of some zooplankton species deepened following the invasion of
Bythotrephes in Lake Michigan (Lehman and Caceres 1993) and Lake Eric (0. E.
Johansson, unpublished data). These patterns are indicative of diel vertical migration, a

plastic behavioral trait commonly observed in zooplankton (Gliwicz 1986; Dodson 1988;

reviewed in DeMeester et al. 1998). Die] vertical migration may reduce the risk of
zooplankton being eaten by Bythotrephes, as this predator is shallow-dwelling and uses
visual cues to detect prey (Muirhead and Sprules 2003), but migration may also come at a
cost to zooplankton caused by inhabiting cooler environments (Loose and Dawidowicz
1994). If zooplankton are migrating in response to Bythotrephes, the resultant NCEs of
Bythotrephes may contribute to the net effect of this predator on zooplankton populations
and its indirect effect of fish, by reducing the availability of zooplankton.

My doctoral research centered on evaluating the importance of NCEs to the net
effect of Bythotrephes on the pelagic food web of the Laurentian Great Lakes. The
approach I took was necessarily very different than those taken by previous studies. To
evaluate NCEs, I relied on a sequence of steps that integrated experimental findings, field
observations, and modeling. First, I tested the phenotypic response of the native prey
(zooplankton) to Bythotrephes and related their response to zooplankton phenotypes
observed in the field. Second, I evaluated the possible consequences of the phenotypic
response of zooplankton on their population grth rate (i.e., NCEs of Bythotrephes) in
the field. Third, I incorporate NCEs into food web models to quantify the contribution of
NCEs to the net effect of Bythotrephes on zooplankton populations and to the net indirect
effect of Bythotrephes on an ecologically and economically important Great Lakes
species, alewife (Alosa pseudoharengus).

My dissertation is divided into 6 chapters, which I summarize below. The
chapters are ordered in a logical progression, from measuring the traits of individual
zooplankton to evaluating the consequences of NCEs food web interactions in the Great

Lakes. Two of the chapters are published in peer-reviewed journals (Chapters 1 and 5),

one chapter is in review for publication (Chapter 4), and three chapters have not yet been
submitted for publication (Chapters 2, 3, and 6) (Appendix A).

Chapter 1 was an initial test to determine if zooplankton respond to Bythotrephes
by modifying their phenotypes. I focused specifically here and in other parts of my
dissertation work on the zooplankton species, Daphnia mendotae, which dominates the
zooplankton community during the summer in Laurentian Great Lakes (Barbiero and
Tuchman 2001). I found that Daphnia exposed to Bythotrephes kairomones (water borne
chemicals extruded by the predator [Lass and Spaak 2003]) migrated downward at
substantial cost to their somatic growth rate. Results further indicated that concentrations
of Bythotrephes kairomones in water taken directly from the field (Lake Michigan) were
high enough to induce behavioral shifts that led to these large reductions in somatic
growth rate.

Results of Chapter 1 indicated that Daphnia mendotae was not naive to risk posed
by Bythotrephes. This finding is counter to the notion that prey may be unable to perceive
and respond adaptively to novel risk posed by invasive predators because they lack
evolutionary experience with the predator. Chapter 2 extends the tests of naiveté to
include other Lake Michigan zooplankton prey and another invasive predator,
Cercopagis pengoi. Cercopagis has been present in the Laurentian Great Lakes for six
years and is in the same phylogenetic family as Bythotrephes, but the relatively small-
sized Cercopagis poses a lesser risk to some zooplankton than the larger Bythotrephes. I
also evaluated the migratory response of zooplankton to four common native predators to
determine whether the migratory response of D. mendotae induced by Bythotrephes was

a generalized response, thereby providing a possible explanation for the lack of naivete’. I

found that zooplankton also responded adaptively to Cercopagis, and that downward
migration is a generalized response induced by native predators that are functionally, but
not phylogenetically similar to Bythotrephes and C ercopagis. Although native prey in
aquatic systems are thought to be prone to naivete' (Cox and Lima 2006), the adaptive
response of zooplankton indicates that nai'veté currently does not play a role in their
interactions with invasive predators.

Chapter 3 examined consequences of such vertical migration on Daphnia grth
rate in the field. Migratory zooplankton experience fluctuations in their environment due
to the multiple environmental gradients associated with the water column. I focused on
two gradients, water temperature and food resources, that have been shown to influence
zooplankton growth and reproduction in other systems (Williamson et al. 1996; Winder
et al. 2003; Kessler and Lampert 2004). In Lake Michigan, these gradients may counter
each other, as the highest concentrations of algae, a primary food resource for
zooplankton, are found at deep chlorophyll maxima (DCM) where water temperature is
relatively cold (Barbiero and Tuchman 2001). Results indicated that variation in water
temperature experienced by migrating Daphnia, but not variation in food resources,
greatly influenced grth rate of Daphnia and demonstrated that inhabiting cooler
environments is the primarily cost of migration.

In order to determine the net effect of Bythotrephes on zooplankton, it was
necessary to also accurately predict the CE of Bythotrephes. Chapter 4 evaluated
predation by Bythotrephes over a natural gradient of light intensity. Light intensity may
be particularly important to Bythotrephes interaction with prey, because Bythotrephes

encounter most prey during the nighttime and Bythotrephes' ability to detect prey

decreases with decreased light intensity (Muirhead and Sprules 2003). Results here
showed that Bythotrephes predation varied among the different light intensities, with
predation not detectable under low light intensity (<1 umol m'2 s"), but increasing with
greater light intensity, and eventually asymptoted under high light intensity (>100 umol
In2 5"). These changes in consumption indicate that Bythotrephes predation is more
sensitive to light than previously thought, a discrepancy that can be explained afier
considering the ability of D. mendotae to detect Bythotrephes' hydromechanic
disturbance. In addition to aiding predictions of the CE of Bythotrephes in the field, these
finding also provide a novel explanation for the tendency of Bythotrephes to invade lakes
of high water clarity.

Chapter 5 combines the findings of the first four chapters with observations from
an intensive field survey to evaluate NCEs of Bythotrephes on zooplankton prey
populations in Lakes Michigan and Eric. Thus, to my knowledge, this chapter represents
the first study that estimated NCEs on prey population growth rate in a natural setting.
The results revealed that NCEs of Bythotrephes contribute substantially to net effect of
Bythotrephes on several prey population growth rates, with NCEs on the same order of
magnitude as or greater (up to 10 fold) than CEs.

Chapter 6 explored how NCEs of Bythotrephes on zooplankton indirectly affected
age-0 alewife by incorporating NCEs observed in Chapter 5 into a food web model.
Model predictions of seasonal population dynamics of Daphnia and consequent growth
of alewife indicated that the impact of Bythotrephes on alewife depended strongly on
whether or not the NCEs of Bythotrephes were included. For example, when only CEs of

Bythotrephes were considered, predicted mean length of alewife entering their first winter

was 110 mm and predicted over-winter survival was 87.1 %. When NCEs were added,
both by including a cost of migration on Daphnia population growth rate and considering
the reduction in overlap between alewife and Daphnia resulting from migration,
predicted mean length of alewife entering winter was 72 mm and predicted over-winter
survival was 19 %. These findings provided an example in a natural system of how
interactions between competitors can be strongly influenced by behavioral responses of
shared prey.

The results of my dissertation research have implications both specific to how
biologists assess the impact of Bythotrephes on zooplankton prey populations and the
Great Lakes pelagic community, and general to predator-prey interactions in nature.
Specifically, my research aids Great Lakes managers in an immediate and important
problem by providing more comprehensive assessment of Bythotrephes impacts.
Although I have focused on indirect interactions between Bythotrephes and alewife
interactions, the approach I took may also be used to identify effects of Bythotrephes on
other species and the indirect impacts of other invasive species. In general, my research
bridged recent knowledge developed in community ecology to large-scale ecological
patterns and process, and addressed critical questions, primarily how important are NCEs
in ecosystems as complex as Lake Michigan’s. The product of this inquiry will hopefully
alert researchers of diverse systems to the potential importance of NCEs and provide a

mechanistic approach that can be applied to both aquatic and terrestrial environments.

CHAPTER ONE

NON-CONSUMPTIVE EFFECT OF THE INVASIVE PREDATOR B Y T HOT REPHES
LONGIMAN US ON DAPHNIA MENDOTAE

ABSTRACT
We evaluated the antipredator behavior of Daphnia mendotae to the invasive invertebrate
predator, Bythotrephes longimanus, and the consequent effect of the predator on prey
growth rate (referred to as a non-consumptive effect of the predator). In a laboratory
experiment, Daphnia in the absence of Bythotrephes kairomones remained in the top,
warmer regions of experimental columns, whereas in the presence of Bythotrephes
kairomones, Daphnia migrated vertically, occupying a middle region by night and a low,
cold region during the day. Over a 4-day experiment, the vertical migration induced by
Bythotrephes caused a 36 % reduction in the somatic growth rate of Daphnia, a level that
is sufficient to have an effect on prey population growth rate. A second laboratory
experiment indicated that concentrations of Bythotrephes kairomones in water taken
directly from the field (Lake Michigan) was high enough to induce behavioral shifts that
led to these large reductions in somatic growth rate. Our results identify a means by
which Bythotrephes has substantial effects on native prey populations other than through

direct consumption.

INTRODUCTION

Understanding the impact of invasive species is one of the most important
challenges facing biologists today (Mack et al. 2000; Byers et al. 2002; Hochberg and

Gotelli 2005). One clear effect of invasive species is predation on native prey; however,

10

potential non-consumptive effects are is less well understood. Non-consumptive effects
occur when prey respond to predators by a modification of their phenotype (e.g.,
behavior, morphology, and life-history) that may reduce predation risk (Lima 1998). In
theory, this phenotypic response increases fitness, but due to tradeoffs, there are costs,
often in the form of reduced growth and fecundity (reviewed in Harvell 1990). Empirical
(Lima 1998; Werner and Peacor 2003; Schmitz et al. 2004) and theoretical studies
(Abrams 1982; Ives and Dobson 1987; Abrams 1993; Peacor and Werner 2004; Bolker et
al. 2003) indicate that such non-consumptive effects can be strong relative to direct
consumption. Invasive predators have been shown to have non-consumptive effects on
native prey through induced phenotypic changes in several systems, including streams
(McIntosh and Townsend 1994), ponds (Shave et al. 1994; Pearl et al. 2003) and
terrestrial communities (Wilder and Rypstra 2004). Thus, whereas it is important to
understand the non-consumptive effect of invasive predators to understand their impacts
fully, our present understanding is derived from a limited number of systems.

Here, we examined non-consumptive effects of an invasive predator,
Bythotrephes longimanus (Leydig), on zooplankton prey in the pelagic zone of Lake
Michigan through a well-known antipredator behavior, vertical migration. Bythotrephes
is an invasive predatory cladoceran that studies suggest has had an impact on the pelagic
ecosystem of the Laurentian Great Lakes (Lehman and Caceres 1993; Vanderploeg et al.
2002; Barbiero and Tuchman 2004), and smaller surrounding lakes (Yan et al. 2001;
Boudreau and Yan 2003). In particular, since its introduction in the early 1980’s,
Bythotrephes has been described as a main factor driving community dynamics in Lake

Michigan (Makarewicz et al. 1995). Although research on the effects of Bythotrephes on

11

its zooplankton prey has focused primarily on direct consumption, field observations in
Lake Michigan indicate that Bythotrephes may induce downward migration of
zooplankton prey (Lehman and Caceres 1993; K. L. Pangle, unpublished), which could
affect prey growth rate. However, alternative processes could be responsible for the
changes in zooplankton vertical distribution including: (1) another factor that is
correlated with Bythotrephes density; (2) greater Bythotrephes predation of the entire .
prey population near the surface; (3) seasonal shifts in clones within a prey species that
exhibit different behavior; or (4) an indirect effect from Bythotrephes affecting some
other environmental condition, such as attracting fish.

We used an experimental approach to examine if Bythotrephes changed the
vertical of Daphnia mendotae (Birge), and any non-consumptive effect on grth rate
caused by such a behavioral modification. Daphnia mendotae is an important food for
Bythotrephes (Lehman et al. 1997; Schulz and Yurista 1999), and Daphnia is model
organism for the study of predator-induced tactics (Lass and Spaak 2003). Daphnia
mendotae has also been shown to respond to Bythotrephes kairomones (i.e., water—borne
chemicals produced by the predator) by growing larger helmets (Bungartz and
Branstrator 2003). In the laboratory, we exposed D. mendotae to Bythotrephes kairomone
over a 5-day period. We hypothesised that D. mendotae would respond by migrating
downward, particularly during the day, because Bythotrephes is a visual predator
(Muirhead and Sprules 2003) that inhabits the epi- and metalimnetic portions of the water
column in Lake Michigan (Lehman and Caceres 1993; K. L. Pangle, unpublished).
Further, we hypothesised that vertical migration into deeper, colder water would incur a

cost to growth. A second laboratory experiment was conducted to link our laboratory

12

results with the concentrations of Bythotrephes kairomones in water collected from Lake

Michigan.

METHODS

We measured the effect of Bythotrephes kairomones on the behavior and grth
rate of Daphnia mendotae in experimental cylinders with a thermal gradient spanning a
similar temperature range as that found in Lake Michigan. Experiment 1 was conducted
to examine the effects of Bythotrephes kairomones on the vertical distribution of Daphnia
and on individual (i.e. somatic) growth. Experiment 2 compared the response of Daphnia
to lake water collected from regions with and without Bythotrephes and to a range of
Bythotrephes kairomone concentrations prepared in the laboratory.

The experimental system followed that of Loose and Dawidowicz (1994). During
the experiments, the vertical position of Daphnia was measured in 80-cm tall, 19-mm
diameter, clear acrylic cylinders that were filled with experimental water and submerged
vertically in a transparent, 300-L aquarium. The aquarium acted as a thermally-stratified
water bath, with water temperatures regulated to 23°C at the surface and 12°C at the
bottom, using an external chiller unit. The cylinders were illuminated by diffused light
from directly above using four, 50W halogen bulbs. Light conditions followed a 14 h
light : 10 h dark regime centred at 12:00, with a photon flux density of 20.0 mol m'2 s'1
at the top of the cylinder and 4.0 pmol rn'2 s'1 at the bottom during midday. During the
transitions from day to night or vice versa, light intensity was gradually increased or
decreased over a 2-h period. Water was pumped into each cylinder using two 24-channel

peristaltic pumps via individual 0.38 mm-diameter tygon tubes at a rate of 13.3 ml h'l,

13

creating a 12-h residence time for the water in each cylinder. Given this rate of water
flow, the supply rate of resources was approximately 20 times greater than the expected
maximum clearance rate of the Daphnia (0.6 ml h'1 individual"; Lurling 2001), and
therefore any effects of Daphnia grazing of food were negligible.

In both experiments, we measured the vertical position of Daphnia by recording
the depth interval occupied. Cylinders were demarcated into 10 intervals total, with two
5-cm intervals at the top, two 5-cm intervals at the bottom, and six lO-cm intervals in the
middle. We used mean depth (cm) and the percentage of individuals in the upper and
lower 5-cm extremes of the cylinder as measures the vertical position of Daphnia.
"Depth" was defined as the midpoint of the depth interval in which Daphnia were
observed.

Daphnia mendotae were collected from Lake Michigan on 12 July 2004 using a
zooplankton net with a closed bucket towed vertically through the entire water column.
Collections occurred approximately 8 km offshore of Muskegon, Michigan, an area
inhabited by Bythotrephes (K. L. Pangle, personal observation). The Daphnia were then
maintained in 4-L vessels in aged Lake Michigan water at 20°C with a 14 h light : 10 h
dark light regime. They were fed Nanochloropsis limnetica (Hibberd) (SAG 18.99,
University of Gottingen, Sag, Germany) at approximately 6 mg C L'l. The Bythotrephes
used in the experiments were collected in a similar fashion at the same location on 6
August and 7 September 2004, and maintained individually in 50-ml glass containers at
the same temperature and light conditions as Daphnia and were fed 5 Daphnia d".

Experiment 1 was conducted from 9 to 13 August 2004. During the experiment,

Daphnia were kept in either a control water treatment or in a Bythotrephes water

14

treatment. Control water was lake water passed through a 0.45-um filter and aged at
room temperature over a 14-d period, a sufficient duration for breakdown of pre-existing
kairomones (Loose et al. 1993). To create Bythotrephes water (density, 5 Bythotrephes L'
l), individual Bythotrephes were incubated in control water in separate 200-m1 containers
for a 24-h period. Fresh Bythotrephes water was created on seven separate occasions over
the course of the experiment, so it could be replaced every 12 hours. Bythotrephes were
not fed during the incubation period. The control water used in the experiment was
maintained under similar condition with no Bythotrephes added. Two hours prior to the
experiment, Bythotrephes were removed from each container, and Bythotrephes water
and control water passed through a 64-pin filter to remove any debris. For food,
Nanochloropsis was added to all experimental water (6 mg L'l). Some of the
experimental water was used to fill respective cylinders, while the remainder was
transferred to reservoirs that supplied the peristaltic pump system. Treatments were
assigned to cylinders using a randomized block design; each of the two treatments was
represented in 10 spatial blocks, with each consecutive two cylinders representing a block
(20 cylinders aligned side by side in total).

Daphnia used in the experiment were neonates, born four to five days beforehand.
This ensured that the organisms were initially of uniform size and developmental state,
and the young age reduced the chance of Daphnia producing offspring during the
experiment. One Daphnia was transferred into each cylinder and allowed to acclimate to
experimental conditions for 1 h before observations began. An additional 20 Daphnia

were picked haphazardly and measured to estimate the initial size.

15

The experiment started at noon and continued for four days. The vertical position
of each Daphnia was determined every 12 hours (nine observations in total), coinciding
with noon and midnight. At the end of the experiment, the Daphnia were removed from
each cylinder and preserved in a sugar-formalin solution. Standard body length of each
Daphnia was later measured under 40x magnification using a drawing tube and digitizer
(Roff and Hopcroft 1986). Standard body length was converted to dry weight using a
standard equation (Dorazio et al. 1987) to determine the absolute weight gain over the
duration of the experiment. In addition, specific growth rate (SGR) in mass was
calculated as:

ln(

W )-ln(W.)
(1.1) SGR= f ’

d

 

where Wf is final dry weight (pg), W,- is initial dry weight (pg), and dis the duration of
the experiment (days).

Experiment 2 was conducted on 8 September 2004. Daphnia were kept in one of
six water treatments including: control water; three concentrations of laboratory prepared
Bythotrephes water, low, medium, and high; lake water not inhabited by Bythotrephes;
and lake water inhabited by Bythotrephes. This design allowed us both to compare the
concentrations of kairomones in Lake Michigan with those created in the laboratory, and
also to evaluate the effect of predator density on vertical migration. Lake water was
collected from depths of the water column where Bythotrephes is known to be most and
least abundant (K. L. Pangle, unpublished). Specifically, 1000 L of lake water was
pumped from l-m and 20-m depths where Bythotrephes was absent (no Bythotrephes

were found) and present (approximately 0.010 Bythotrephes L"), respectively, at 16:00

16

on 7 September 2004 at the same site where experimental Daphnia and Bythotrephes
were collected. At the time of sampling, the Lake Michigan water column was stratified,
with a thermocline at approximately 20-m. During the period between water collection
and the start of the experiment, lake water was maintained in the dark at 4°C to reduce
kairomone breakdown. The control water and water prepared in the laboratory with a
high density of Bythotrephes (5 L") was made in a similar fashion as Experiment 1.
Laboratory prepared water with low and medium Bythotrephes density (0.05 and 0.5 L],
respectively) was made through serial dilutions of the "high density" prepared water. For
food, Nanochloropsis (6 mg L'l) was added to all experimental water. In contrast to
Experiment 1, six replicate cylinders were designated to each treatment, and each
cylinder received six D. mendotae. Treatments were assigned to cylinders using a
randomized block design; each of the six treatments was represented in six spatial blocks,
with each consecutive six cylinders representing a block (36 cylinders aligned side by
side in total). Five observations were taken from 10:00 to 14:00 in 1-h intervals. All other
experimental conditions were as described in Experiment 1.

In Experiment 1, a repeated-measures analysis of variance (ANOVA) was used to
evaluate the effect of Bythotrephes kairomone on Daphnia vertical position during the
day and night over the duration of the experiment. In addition, the secondary effects of
Bythotrephes kairomone on standard body length and specific grth rate of Daphnia
were evaluated using a blocked one-way ANOVA. In Experiment 2, the behavioral
responses were averaged over all Daphnia within each cylinder over the duration of the
experiment. The behavioral response of Daphnia to each treatment was analyzed using a

blocked one-way ANOVA. When significant differences among treatments were

17

detected, a Tukey pair-wise comparison was used to separate the means. All percentage

data were arc-sin transformed prior to analysis (Zar 1999).

RESULTS

Daphnia responded to Bythotrephes kairomones by migrating downwards into
cooler water. In Experiment 1, Daphnia remained high in the experimental columns by
both day and night in the absence of Bythotrephes kairomones but, when in the
Bythotrephes water, the Daphnia were significantly deeper (F 1,1 57 = 91 .06;p < 0.01),
particularly during the day (significant interaction between kairomones and photoperiod
[F 1,157 = 16.93; p < 0.01]). Thus, in the presence of Bythotrephes, Daphnia displayed a
diel vertical migration, but not in the predator's absence (Figure 1.1). The percentage of
individuals in the upper 5-cm section of the cylinder was also significantly lower for
Daphnia in the Bythotrephes water than for those in the control water (F 1,157 = 37.50, p <
0.01), with no significant difference between day and night (F 1,1 57 = 1.97, p = 0.16).
Further, the percentage of individuals in the lowest 5-cm section of the cylinder was
significantly greater for the Daphnia in the Bythotrephes water than for those in the
control water during‘the day, but not at night, leading to a significant interaction (F 1,157 =
5.20, p = 0.02).

Bythotrephes kairomones also affected Daphnia growth. Over the duration of the
experiment, the mean weight gained by Daphnia in the Bythotrephes water (1.2 i 0.2 pg)
was less than one half that in the control water (2.7 d: 0.2 pg, F 1,13= 22.34; p < 0.01).
Daphnia mean specific growth rate in Bythotrephes water, 29.5 d: 3.3 % d}, was 36 %

lower than in the control water, 45.7 d: 1.8 % cl'I (F 1,13: 13.02; p < 0.01). In Experiment

18

.323 mm

av :33 SSE 2: Bumbag 3&8 weanénwt 22. .mEEBE 80% N Seton 05 H8 085 26: 8 68388 :03 96: $3.85 59m
N no“ 05 8m mowficoouom :32 A3285 Snow 338%8 :05 Mo “5828 ofi um $5...-» 2: :o 3:358 upon 96: 3503 3283
:33 some “a SEESQ go Amado mm “3 owficoeoa :38 2: “comeback maoudnmbmmu 3:252.“ 985-co— ofi. 6:080:8— nonmmhofiam
£07286 some 53> cum 2; mus—86 38v 3923 53>» 83 cows 8 3898 ocuohaofi .Q «0 Amie was: 2th 28809:“:

23 A38 28: an: nice we 830:3 a ma 332m .5363 Bozo? 3v Ema E8 A3 b5 ._ Eofitoqu .«o $38M .: 2:me

 

g 3295 came a 585a g 5225 same a 58.8
8 8 om o 8 2. 8 o
O.N_\ _ . _ . _ . _ _ om
3.39853 O
.8280 .
3T . I 8
M
8
. a
1 G
8
w m
m 2:1 - - ov N
e O
m w
m I.\
0..
(x o ONI @ . i ON

 

 

 

 

E E
08 o
£92 :3

 

19

2, D. mendotae responded similarly to a gradient of Bythotrephes kairomone
concentration in both water collected directly from the lake and prepared in the
laboratory. Mean depth of Daphnia differed significantly among treatments (F 3,27 =
45.32; p < 0.01), so pair-wise comparisons were made between treatments (denoted
significant below if p was less than 0.05). Relative to the control, mean depth was not
significantly different for Daphnia in lake water not inhabited by Bythotrephes, but was
significantly deeper for Daphnia in lake water inhabited by Bythotrephes (Figure 1.2a).
Likewise, the vertical position of Daphnia was significantly deeper as Bythotrephes
kairomone concentration increased in laboratory-prepared treatments. The behavior of
Daphnia in lake water not inhabited by Bythotrephes closely resembled those in control
water, whereas the behavior of Daphnia in lake water inhabited by Bythotrephes was
similar to that in water with Bythotrephes kairomones that had been prepared in the
laboratory (Figure 1.2a). Significant differences were also found among treatment in the
percentage of individuals in the extreme sections of the cylinder (Figures 1.2b & 0). Pair-
wise comparisons between treatments indicated that Daphm'a in lake water not inhabited
by Bythotrephes and in control water were found more frequently in the upper 5-cm
section than Daphnia in water from Bythotrephes kairomone treatments (F 3,27 = 12.50; p
< 0.01). Further, a significantly greater percentage of Daphnia were observed in the
lower 5-cm extreme of the cylinder in water from Bythotrephes kairomone treatments
than those in lake water not inhabited by Bythotrephes and control water (F 3,27 = 11.32; p

< 0.01).

20

 

(a)

20‘

40-

Depth (cm)
l-O-l ‘<
l—O-l
é

l-O-lx

60‘

 

30 - (b)

20-

Percentage In the
upper extreme (%)

 

30 — (C)

204

yz yz
.zzfl.$rfi

Control Absent Present Low Medium High

Bythotrephes in Bythotrephes in
lake water laboratory water

Percentage in the
lower extreme (%)

 

 

 

Figure 1.2. Mean depth (a) and the percentage of D. mendotae in the upper and lower 5-
cm extremes (b and c, respectively) in Experiment 2. Treatments included: control water;
lake water not inhabited by Bythotrephes; lake water inhabited by Bythotrephes
(approximately 0.010 Bythotrephes L"); low Bythotrephes-density laboratory-prepared
water (0.05 Bythotre hes L'l); medium Bythotrephes-density laboratory-prepared water
(0.5 Bythotrephes L' ); and high Bythotrephes-density laboratory-prepared water (five
Bythotrephes L'l). Values reported are treatment means (:h SE). Means with different
lowercase letters were significantly different as determined by ANOVA and Tukey’s
HSD tests (p < 0.05).

21

DISCUSSION

Our results support the hypothesis that the presence of Bythotrephes induces a
behavioral response by Daphnia. In both laboratory experiments, Daphnia exposed to
Bythotrephes kairomones during the day were found in lower, cooler regions of the
experimental columns, while Daphnia exposed to water with no kairomones inhabited
upper, warmer regions. The difference between treatments lessened in magnitude when
light intensity decreased, as seen in Experiment 1, in which Daphnia exposed to
Bythotrephes kairomones migrated upward during the night, while the vertical
distribution of Daphnia in the control treatment remained unchanged. Experiment 2 also
showed that Daphnia responded in similar ways to lake water inhabited by Bythotrephes,
but not without, suggesting that the concentration of kairomones in the field is sufficient
concentrations to illicit a behavioral response by Daphnia.

The alternating day-night movement of Daphnia observed in our study (i.e., diel
vertical migration), normally associated with a response to vertebrate planktivores (Stich
and Lampert 1981; Gliwicz 1986; Loose and Dawidowicz 1994; De Meester et al. 1998),
could optimize the fitness of Daphnia when exposed to predation risk from Bythotrephes.
Bythotrephes, unlike many aquatic invertebrate planktivores (Riessen et al. 1988; Roche
1990), uses visual cues to detect prey and may be five times more effective as a predator
in the photic zone of the water column than in the aphotic zone (Muirhead and Sprules
2003). Subsequently, inhabiting deeper, darker regions of the water column during the
day would greatly reduce predation risk for Daphnia. At night, Daphnia can increase
growth and developmental rates by migrating upwards into warmer near-surface regions

when the risk from Bythotrephes is lower.

22

Our results are consistent with changes in Daphnia vertical distribution observed
in pelagic systems invaded by Bythotrephes. For example, in Lake Michigan, a
downward shift of D. mendotae in the water column corresponded with the invasion of
Bythotrephes (Lehman and Caceres 1993), and there is a strong positive relationship
between the percentage of D. mendotae in the hypolimnion and Bythotrephes abundance
that occurs seasonally in Lake Michigan (K. L. Pangle, unpublished). In Harp Lake,
Ontario, D. mendotae maintained a deeper vertical distribution than that of Bythotrephes
during the day, but overlapped with the predator at night (Dumitru et al. 2001; Muirhead
and Sprules 2003). In each case, the vertical migration induced by Bythotrephes could be
responsible for the vertical distribution observed in the field. Note, however, that there
are a number of possible alternative mechanisms for the observed field patterns (listed in
the Introduction), and therefore this laboratory study makes clear the potential
contribution of the behavioral response.

We predicted that predator-induced diel vertical migration would be associated
with a cost, as is general to anti-predator tactics (reviewed in Harvell 1990; Lima 1998).
In Experiment 1, the vertical migration of Daphnia to Bythotrephes kairomone led to a 36
% reduction in somatic growth. This reduction could have been caused by several
mechanisms including inhabiting cooler water (Loose and Dawidowicz 1994) and
allocating more energy to early reproduction and morphological defenses relative to
somatic growth (Tollrian 1995).We evaluated how these possibilities contributed to the
non-consumptive effects on Daphnia in Experiment 1.We calculated a predicted somatic
growth rate (SGR) for each individual (given their thermal experience during the

experiment) using the equation:

23

(1.2) SGR = —0.0009 T2 + 0.0476 T- 0.1882

where T is water temperature (°C). This equation was derived by fitting a nonlinear model
to published data (Hall 1964) that reported D. mendotae growth rate under saturated food
resource conditions over a wide range of temperatures (11 to 25°C). The predicted
reduction in Daphnia growth in the Bythotrephes treatment was close to the actual
reduction (Figure 1.3). This suggests that migratory behavior into cooler water, and not
modifications to morphology and life history, were principally responsible for the non-
consumptive effect on growth rate.

For Daphnia species, somatic grth can be used as an approximation of fitness,
as it is known to correlate strongly with population grth rate (Lampert and
Trubetskova 1996); thus, our results show a means by which Bythotrephes affects prey
populations other than by direct consumption (i.e., non-consumptive effect). These so-
called non-consumptive effects are predicted theoretically to play a critical role in
predator-prey interactions (Abrams 1982; Ives and Dobson 1987; Abrams 1993). Indeed,
using simple models, Peacor and Werner (2004) showed that predator-induced reductions
in growth rates on the order observed here (36 %) are large enough to contribute
substantially to the net effect of the predator on prey fitness. Empirical studies performed
in mesocosms have also shown that the magnitude of non-consumptive effects on prey
population grth rates can be as large as or larger than that of lethal effects (Nelson et
al. 2004; Boeing et al. 2005). Note that, in addition, such non-consumptive effects may
extend indirectly to affect prey food resources and competitors (reviewed in Abrams
1995; Bolker et al. 2003; Werner and Peacor 2003; Schmitz et al. 2004) via trait-

mediated indirect interactions, further contributing to the net predator-prey interaction.

24

50

 

 

[:1 Observed
Predicted
g 40 ~
o\°
.9
e 1L
.:
E
9
0’ 20 -
o
E
o
a)
c.
U) 10 -
0

 

 

 

 

 

 

 

Control Bythotrephes

Treatments

Figure 1.3. Observed (open bars) and predicted (using a grth model, hatched bars)
specific grth rate of D. mendotae in Experiment 1 exposed to aged lake water without
and with Bythotrephes kairomone. Values reported are treatment means (1: SE).

25

These results, combined with our study, suggest that the induction of diel vertical
migration may be as important to Daphnia population dynamics as changes caused by the
direct consumption of Daphnia by Bythotrephes.

In conclusion, this study demonstrated the behavioral response of D. mendotae to
Bythotrephes and a consequent non-consumptive effect on growth. Our findings may be
applicable to other systems, as Bythotrephes density used in the experiments were not
only typical of Lake Michigan (Pothoven et al. 2003), but also of those observed in Lakes
Erie and Huron (Barbiero and Tuchman 2004), smaller Canadian inland lakes (Boudreau
and Yan 2003), and lakes in the native range of Bythotrephes (Manca et al. 2000; Palmer
et al. 2001). Moreover, our study emphasizes the complex nature of the impact of
invasive species as well as provides an avenue for research into better understanding
Lake Michigan food web dynamics. What is the relative magnitude of non-consumptive
effects compared to lethal effects and how are they transmitted through the food web via
trait-mediated indirect interactions? Further, how do non-consumptive effects occur
under such short-term evolutionary periods? Addressing these questions could shed light

into both the ecology and evolution of invasive species.

26

CHAPTER 2

LACK OF NAIVETE IN LAKE MICHIGAN ZOOPLANKTON PREY TO RISK
POSED BY INVASIVE PREDATORS

ABSTRACT
Prey have evolved traits that enable them to perceive and respond adaptively to predation
risk. However, prey may be na'r’ve towards novel risk posed by invasive predators, and
prey naivete’ could profoundly influence the nature of invasive-predator-prey interactions.
Here, using a series of laboratory experiments, we tested the na'r'veté of three Lake
Michigan zooplankton species by evaluating their behavioral response (i.e., vertical
migration) to the presence of two invasive invertebrate predators, Bythotrephes
longimanus and C ercopagis pengoi. Bythotrephes can consume all three zooplankton
species used in the study, but the smaller Cercopagis is hindered by the large-sized prey,
and subsequently posed a real risk to only the smallest one of the zooplankton species.
We also evaluated the behavioral response of zooplankton in the presence of four native
predator species, Percaflavescens, Leptodora kindtii, Polyphemus pediculus, and Mysis
relicta, which vary in their phylogenetic and functional similarity to the invasive
predators. The results indicated that prey were able to distinguish between the presence
and absence of Bythotrephes and C ercopagis and responded behaviorally by migrating
downward only when it was adaptive. Thus, zooplankton appeared not be na'r've towards
the risk posed by these invasive predators. Downward migration was also induced by a
functionally similar native predator, suggesting that this behavioral trait is an exadaption

that is sufficiently general to be co-opted to avoid invasive predators. Although aquatic

27

systems are thought to be prone to na'r'veté, the results suggest that na'r'veté does not

contribute to current impact of Bythotrephes and Cercopagis in Lake Michigan.

INTRODUCTION

Prey can evolve traits that reduce their risk of being consumed by predators.
However, prey may lack appropriate traits to reduce novel risk posed by invasive
predators due to lack of co-evolution, a condition referred to as prey na'r'veté. Na'r'veté may
leave prey extraordinarily vulnerable to the invader, and has been linked to enhanced
growth of invasive predator populations, extirpation of prey populations, and an overall
destabilization of invaded systems (Fritts and Rodda 1998; Blackburn et al. 2004; Cox
and Lima 2006; Salo et al. 2007). However, lack of co-evolution does not guarantee an
easy meal for invasive predators. Prey may have evolved antipredator traits through
experience with native predator (exadaptations) that are sufficiently general to be co-
opted to thwart invasive predators (Wilder and Rypstra 2004). Due to the potentially
strong selection pressures posed by invasive predators, native prey may also adapt traits
or learn strategies that minimize novel risk at a rate that outpaces extinction (Griffin
2004; Strauss et al. 2006).

Whether or not prey na'r'veté occurs may be dependent on the characteristics of the
invaded ecosystem, and variation in na'r'veté may provide insights into why some
ecosystem types tend to be more impacted by invasive predators than others. A primary
factor driving differences among ecosystem types is insularity. In insular ecosystems,
prey tend to adapt defenses to combat local predator archetypes but lack evolutionary

experience and thus possibly appropriate traits to respond to predator archetypes outside

28

of their system. Classic and extreme examples of such insular ecosystems are oceanic
island ecosystems that support prey populations with few or no native predators. The lack
of evolutionary experience with predators can render island prey highly susceptible to
invaders (Case and Bolger 1991; Fritts and Rodda 1998), and may be an underlying
reason for the greater impact invasive predators have on native prey of islands relative to
mainlands (Simberloff 1995; Blackburn et al. 2004; Salo et al. 2007). Like islands, lake
ecosystems may isolate prey populations due to surrounding land barriers that hinder
dispersal among systems, and, like islands, the insularity of lakes may foster na‘r‘veté in
lake prey species to invasive predators (Cox and Lima 2006). Although prey na'r'veté in
aquatic systems is not yet well understood, the notion that lake prey are generally na'r've
could provide powerful insights into the nature of the impact invasive predators have in
lakes, as well as explain broader patterns among invaded ecosystems. For example, Cox
and Lima (2006) recently proposed that differential na'r’veté is an underlying cause for the
greater impact of invasive species observed in aquatic systems relative to terrestrial
systems (Vermeij 1991; Ricciardi and Rasmussen 1999; Bruno et al. 2005).

In our study, we evaluated the na'r'veté of Lake Michigan zooplankton species
towards invasive planktivores. Previous laboratory and field studies indicate that the
invasive predatory cladoceran, Bythotrephes longimanus, induce downward migration in
one Lake Michigan zooplankton species, Daphnia mendotae (Lehman and Caceres 1993;
Chapter 1; Chapter 5). In this case, vertical migration is a plastic behavioral trait that can
be used by Daphnia mendotae to reduce spatial overlap with the shallow-dwelling
Bythotrephes during the daylight hours during which Bythotrephes pose the greatest risk.

Bythotrephes has only been present in Lake Michigan for about twenty years and belongs

29

to a phylogenetic family (Cercopagidae) previously unrepresented in North America
(Dodson and Frey 2001), a characteristic generally used to designate novel predator
archetypes (Cox and Lima 2006). Induction of vertical migration suggests that D.
mendotae is not naive to Bythotrephes, a result contrary to the prediction of na'r'veté in
lake prey.

The goals of our study were two-fold. First, we wanted to determine whether the
migratory response of D. mendotae induced by Bythotrephes was a general response
induced by native Lake Michigan planktivores, thereby providing a possible explanation
for the lack of na‘r’veté in D. mendotae. Second, we wanted to expand tests of na'r’veté to
other Lake Michigan zooplankton species and to another shallow-dwelling invasive
planktivore, Cercopagis pengoi. C ercopagis has been present in Lake Michigan for six
years and is in the same phylogenetic family as Bythotrephes, but the relatively small-
sized C ercopagis poses a lower risk to some zooplankton than the larger Bythotrephes.
Understanding the response (or lack of response) Bythotrephes evoke in native
zooplankton relative to Cercopagis and native predators can provide insights into the
nature of the impact of Bythotrephes and Cercopagis in Lake Michigan, as well as

potentially alter expectations regarding the generality of na'r'veté in aquatic systems.

METHODS
We measured the migratory response of Lake Michigan zooplankton species
using a series of laboratory experiments. To address the first goal, we compared the
response of Daphnia mendotae induced by Bythotrephes to the responses induced by four

native predators to determine if any of these predators also induced downward migration.

30

We chose a set of native predators that encompassed the primary predator archetypes
found in Lake Michigan. They included two predatory cladocerans, Leptodora kindtii and
Polyphemus pediculus, a freshwater shrimp Mysis relicta and a juvenile fish Perca
flavescens (Table 2.1).

The native predators varied in the phylogenetic and functional similarity with
Bythotrephes, thus providing a means to evaluate if downward migration was general to
either phylogeny or functional characteristics of the predator. Leptodora and Polyphemus
are the most similar to Bythotrephes phylogenetically, all belonging to the suborder
Cladocera (Dodson and Frey 2001 ). If the lack of naiveté of Daphnia towards
Bythotrephes is linked to phylogenetic relatedness of Bythotrephes to native predators,
we would expect that Leptodora and Polyphemus would induce a downward migration in
Daphnia. However, Leptodora and Polyphemus are hindered by the relatively large size
of Daphnia (Herzig and Auer 1990; Packard 2001), a characteristic that makes them
functionally different than Bythotrephes (Schulz and Yurista 1998). Subsequently,
vertical migration may be unnecessary for D. mendotae when facing Leptodora and
Polyphemus, or D. mendotae may modify other traits that may be more adaptive for this
risk than vertical migration. For example, the presence of Leptodora is known to induce
enlarged helmets in D. mendotae (Tanner and Branstrator 2006). In contrast to Leptodora
and Polyphemus, juvenile Perca is phylogenetically disparate, but fimctionally similar, to
Bythotrephes. Juvenile Perca are not hindered by the relatively large size of Daphnia
(F ulford et al. 2006) and uses visual cues to detect prey (Jansen and Mackay 1992), as

does Bythotrephes (Muirhead and Sprules 2003). Further, although migratory response of

31

A38 0052388 .2808 .4 .v: 880:8 0003
30058808 05 5 88 >08 0080:8808 000:3 838:2 00—04 5 0:0. 05 00 80502080 80 092020 80%-N :0 0003 00: 0830:...

32

 

 

8000 We 800282 600580202 .0080? 038: 80:? MBSN
30:05 56 8000020 80803085 4.0080855N 0300: 833308 8E0E30K
30:20 .86 2000005 60803085 80885.3 0380: N383 503083
30:05 3 88.850008 .mwb0888< 430280 030.00 n=0000>0§00x0k
30:05 56 00000030 ”00805085 .880§ 03002: 30028 0.808.209
30:05. md 8000020 8083885 9008028.?V 03002: 8:05.538». 803.0.cofixm

0.00: 8503 059 was 08m 00080 .0020 .8138 EmtO .008: 2.503%

 

888890 05 E 008 8002808 038E: 08 0300.: 05 .«0 0058888 08 0082 ._.N 03081

D. mendotae, to our knowledge, has never been tested using the presence of juvenile
Perca, other fish species have induced downward migration in other Daphnia species
(e.g., Loose and Dawidowicz 1994). If the lack of naiveté of Daphnia towards
Bythotrephes is linked to functional relatedness of Bythotrephes to native predators, we
would expect that Perca would induce a downward migration in Daphnia. Mysis relicta
is both phylogenetically and functionally disparate to Bythotrephes, and served as a
control in our study. Mysis are only related to Bythotrephes at the subphylum level, and,
unlike all of the other predators in this study, Mysis are deep-dwelling, thus downward
migration by Daphnia in response to Mysis presence would actually increase spatial
overlap with Mysis. If the lack of na'l’veté of Daphnia towards Bythotrephes is linked to
either functional or phylogenetic relatedness of Bythotrephes to native predators, we
would expect that Mysis would not induce a downward migration in Daphnia.

To address our second goal, we tested the nai'veté of Daphnia mendotae and two
other native zooplankton, Daphnia pulicaria and Bosmina longirostris, to the other
invasive predator, Cercopagis. Because of its small size, Cercopagis only poses a
substantial risk to Bosmina and not to either Daphnia species. Thus, in this case, a
response by the Daphnia species to Cercopagis would be unnecessary and an indication
of nai'veté. Below we first provide our general experimental methodology and then

describe the details of each experiment.

33

Methodology general to all experiments:

We measured migratory response of zooplankton by maintaining them in vertical
columns and observing their behavior in the presence and absence of predators. The
columns were clear acrylic, 80-cm tall and 19-mm diameter, and submerged in a
transparent, 300-L aquarium. The aquarium acted as a thermally stratified water bath,
with water temperatures regulated to 23°C at the surface and 12°C at the bottom, using an
external chiller unit. The cylinders were illuminated by diffused light from directly above
using four, 50 W halogen bulbs. The lighting created a light gradient in the columns, in
which photon flux density was 20.0 umol m'2 s" at the top and 4.0 umol m‘2 s" at the
bottom. We mimicked the presence of a predator in the columns by using predator
kairomones. Kairomones are waterborne chemicals that are released by predators (Lass
and Spaak 2003). Zooplankton can detect kairomones and use them to perceive and
differentiate different types of risk (Dodson 1988; Laforsch and Tollrian 2004). In our
study, zooplankton were exposed to kairomones by filling the columns with water that
had been incubated with the predators over the 24 hrs prior to the start of the
experiments.

Experiments lasted 4 to 4.5 hours and zooplankton behavior during this period
was quantified by recording their location in the columns. Columns were demarcated into
10 intervals, with two 5-cm intervals at the top, two 5-cm intervals at the bottom, and six
lO-cm intervals in the middle. We used the average depth (cm) of individuals within each
column averaged over all observations as measures of the vertical position of
zooplankton. ‘Depth’ was defined as the midpoint of the depth interval in which

individuals were observed. Because vertical movement of zooplankton was constrained

34

by the water's surface and the bottom of the columns, zooplankton evading predation risk
may collect at the extremes of the columns. Thus we also used the percentage of
individuals in the upper and lower 5-cm extremes of each column as a second measure of
the vertical position of zooplankton, which was calculated based on all individuals within
each column and averaged over all observation periods.

All zooplankton prey species were collected from Lake Michigan using a 64-um-
mesh zooplankton net with a closed bucket towed vertically through the entire water
column. Collections occurred approximately eight km offshore of Muskegon, Michigan.
Prior to the experiments, zooplankton were maintained in 4—L vessels in aged Lake
Michigan water at 20°C with a 14 h light : 10 h dark light regime. They were fed
Nanochloropsis limnetica (SAG 18.99, University of Gottingen, Sag, Germany) at
approximately 6 mg C L'l. Five of the six predators (all but juvenile Perca) were
collected in a similar fashion at multiple sites offshore of Muskegon, Michigan, and
immediately used for production of kairomone water. Specific collection sites for the
predators varied between experiments, so we provide details of collection for each
experiment separately. The sixth type of predator, juvenile Perca, was collected by hook-
and—line from nearshore areas of Lake Michigan and maintained in a 50-1 glass aquaria,
and fed a mixture of zooplankton collected from Lake Michigan prior to their use for
production of kairomone water.

Water used to fill the experimental columns was lake water passed through a
0.45-lm filter and aged at room temperature over a 14-day period, a sufficient duration
for breakdown of pre-existing kairomones (Loose et al. 1993). Predators were not fed

while they were incubated in the experimental water, so that kairomones extruded into

35

water and used for the experiments were associated with prey they had consumed in the
lake prior to collection, or in the case of Perca, prey they had consumed from the mixture
of zooplankton collected from Lake Michigan. Some water received no planktivores and
acted as a control. All experimental water received Nanochloropsis algae (6 mg C L") to

provide zooplankton with a saturated level of food resource during the experiment.

Methodology specific to each experiment:

In the first experiment, Daphnia mendotae was exposed to a control treatment (no
kairomones) and five different kairomone treatments created using Bythotrephes,
C ercopagis, Leptodora, Polyphemus, and Percaflavescens. Invertebrate predators were
collected in two different locations in Lake Michigan. Bythotrephes and Polyphemus
were collected from the same site as D. mendotae. During the collection event, we did
observe Cercopagis and Leptodora in the samples, but in very low numbers.
Subsequently, Cercopagis and Leptodora used in the experiment were collected from a
nearshore site where D. mendotae was still present. Because predator size varied by
orders-of-magnitude (Table 2.1),we used different densities (number L'l) of each
planktivore for the production of kairomone water so that their biomass (mg L") was
approximately equal. Experimental planktivore densities were 0.01 individuals L" for
Perca, 1 individual L'1 for Bythotrephes, and 10 individuals L'1 for Cercopagis,
Leptodora and Polyphemus. Although the relationship between body mass and
kairomone production is unknown, here we assumed direct proportionality as a first
approximation. To create kairomone water for invertebrate planktivores, ten individuals

of each planktivore species were transferred into separate 25-ml vials filled with aged

36

lake water and allowed to incubate at 20°C. To create Perca kairomone water, four fish
were transferred into a 4-L beaker filled with aged lake water and the water was
incubated under similar conditions. Afterwards, kairomone waters were passed through a
64-micron filter and then diluted with aged lake water to achieve experimental kairomone
concentrations. For the experiment, six replicate cylinders were assigned to each
treatment (36 cylinders total), and three D. mendotae (mean standard length, 1 mm) were
transferred into each cylinder. Vertical position of the zooplankton was observed hourly
over 4-h period (five observations total).

In Experiment 2, the same prey species, Daphnia mendotae, was exposed to a
control treatment and two different kairomone treatments created using Bythotrephes and
Mysis. Both predators were collected from the same site as D. mendotae. Experimental
planktivore densities were both five individuals L". To produce Mysis kairomone water,
10 Mysis were transferred into a 2-L beaker filled with aged lake water and incubated at
6°C. Production of Bythotrephes kairomone water was the same as in Experiment 1. For
the experiment, 12 replicate cylinders were assigned to each treatment (36 cylinders
total), and three D. mendotae (mean standard length, 1 mm) were transferred into each
cylinder. Vertical position of the zooplankton was observed in 30-min intervals over 4.5-
h period (10 observations total).

In the Experiment 3, Bosmina longirostris and Daphnia pulicaria were exposed to
a control treatment and two different kairomone treatments created using Cercopagis and
Perca. Production of kairomone waters and the planktivore densities was the same as in

Experiment 1. Six replicate cylinders were designated to each treatment (36 tubes total)

37

and three individuals were transferred into each cylinder. Vertical position was recorded

every 1 hr over a 4-h period (five observations total).

Statistical Analyses

Prior to analyses of treatment effects, we tested assumptions of analysis—of-
variance (ANOVA) for our response variables and evaluated potential confounding
factors that occurred during the experiments. Assumptions of normality and homogeneity
of variance were tested using the Wilks-Shapiro Test and the Levene Test, respectively.
One variable, mean depth, met assumptions in all experiments, while the other two
variables, the percentage of individuals in the upper and lower column extremes, violated
assumptions in all experiments, even after arc-sine transformation.

Prior to analyses of treatment effects, we also evaluated potential confounding
factors that occurred during the experiments. One factor was the change in the number of
zooplankton in some columns due to birth of offspring or death. In addition, some
zooplankton adhered to the surface or to air bubbles on the side of the cylinders and were
not included in observations. We tested whether the final number of individuals in a
column affected our behavioral variables in each experiment using ordinary linear
regression, and found that it was never a significant factor (all ps > 0.2). Another
potentially confounding factor occurred in Experiment 3, in which the zooplankter,
Bosmina longirostris, was difficult to see in the columns due to its small size. For two
columns, one replicate in the Cercopagis treatment and one replicate in the Perca
treatment, all individuals could not be accounted for so these replicates were removed

from the analysis. In addition, for some columns, not all individuals could be accounted

38

for during a particular observation, so these observation periods were not included in the
average of the behavioral variables. We tested whether the number of observations for a
given column during the experiment affected our behavioral variables using ordinary
linear regression, and found that it was never a significant factor (p > 0.1).

Treatment effects were analyzed using ANOVA for mean depth and
nonparametric methods for the percentage of individuals at extremes. For Experiments 1
and 2,we tested for significant differences among treatments using one-way ANOVAs
and Kruskal-Wallis tests (Zar 1999). When significant differences among treatments
were detected, we used parametric and nonparametric Dunnett's tests to identify the
predator-kairomone treatments that significantly differed from the control treatment (Zar
1999). For Experiment 3,we used a full factorial ANOVA and Kruskal-Wallis tests to
evaluate the effects of kairomones, prey species, and their interaction. We also evaluated

the simple main effects of species within each kairomone treatment (Kuehl 2000).

RESULTS
In Experiment 1, the migratory response of Daphnia mendotae varied among the
different predator kairomone treatments. Mean depth of D. mendotae in Bythotrephes and
Perca kairomone treatment columns was significantly deeper than those in the control
treatment and in kairomone treatments of Cercopagis, Leptodora, and Polyphemus (F 5,30
= 7.66, p < 0.01; Figure 2.1a). The latter kairomone treatments did not significantly differ
in mean depth from the control treatment. The percentage of D. mendotae at the upper

extreme of the column in Bythotrephes and Perca kairomone treatments was also

39

 

(a)

20- i E E i

Depth (cm)

60-

80-

30 - (b)

 

20-

Percentage In the
upper extreme (%)

It

 

 

30 - (C)

N
O
A

lower extreme (%)
8
1

Percentage in the

 

 

 

Y I

Control CP BL PP LK

 

 

Invasive predators Native predators

Treatments

Figure 2.1. Mean depth (panel a) and the percentage of D. mendotae in the upper and
lower S-cm extremes (panels b and c, respectively) in Experiment 1. Values reported are
treatment means (:t SE). Treatments included: control water or water with kairomones
from the invasive predators Cercopagis (CP) or Bythotrephes (BL) or from the native
predators, Polyphemus (PP), Leptodora (LK), or Perca (PF). Asterisks indicate predator
treatments that were significantly different (p < 0.05) from the control.

40

significantly lower than that for the control, while other treatments did not differ from the

control ( 125 = 13.3, p = 0.02; Figure 2.1b). The percentage of individuals at the lower

extreme of the column did not significantly vary among treatments ( 125 = 9.06, p = 0.11;

Figure 2.1c)

In Experiment 2, D. mendotae responded to both kairomone treatments, but in
opposite directions. Mean depth varied significantly between all treatments, with the
shallowest value for the Mysis treatment, the middle value for the control treatment, and
the deepest value for the Bythotrephes treatments (F 2, 3 3 = 23.26, p < 0.01; Figure 2.2a).
The percentage of individuals at the upper extreme of the column also varied

significantly among all treatments, with 12 % and no Daphnia found near the surface in

the Mysis and Bythotrephes treatments, respectively ( 122 = 18.54, p < 0.01; Figure 2.2b).

The percentage of individuals at the lower extreme of the column was significantly

greater in the Bythotrephes treatment than in Mysis treatment, but neither differed

significantly from the control ( ,‘(22 = 6.14, p = 0.05; Figure 2.2c).

In Experiment 3, Daphnia pulicaria showed no response to kairomones of the
invasive predator, Cercopagis, but migrated downward in response to kairomones of the
native predator, Perca. Mean depth of D. pulicaria was significantly deeper for the Perca
treatment than the control and did not differ between Cercopagis and control treatments
(F 2,27 = 6.52, p = 0.01; Figure 2.3b). The percentage of D. pulicaria at the upper and
lower extremes of the column did not vary significantly among treatments (all ps > 0.05).
In contrast to D. pulicaria, the smaller prey species, Bosmina Iongirostris, migrated

downward in response to kairomones of both the invasive and native predator. Mean

41

 

(a)

x-

20- I

‘-

40— i

Depth (cm)

60a

80a

 

30 _ (b)

20a

 

l~—-la-

10-

Percentage In the
upper extreme (%)

 

 

 

30_ (C)

20a

10~

Percentage in the
lower extreme (%)

E l
o 'k

I l l

 

 

 

 

 

 

Control Mysis Bythotrephes

Treatments

Figure 2.2. Mean depth (panel a) and the percentage of D. mendotae in the upper and
lower 5—cm extremes (panels b and c, respectively) in Experiment 2. Values reported are
treatment means (:t SE). Treatments included: control water or water with kairomones
from the native, deep-dwelling predator, Mysis, or the invasive, shallow-dwelling
predator, Bythotrephes. Asterisks indicate treatments that were significantly different (p <
0.05) from the control.

42

 

(a) (b)

20< f g {

 

:5; .
I 40 < * c
8
o l

60 .

80 «

30 . (C) (d)

20*

Percentage m the
upper extreme (%)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

O . . ,
e
A 30 .< > (t)
2 as.
c m
._ E *
8, e 2°l
B 3?
g a)
L
I— a) 10 4
as: l l
0 . l . l . E l l . .
Control CP PF Control CP PF
Bosmina Iongirostris Daphnia pulicaria

Figure 2.3. Mean depth (panel a, b) and the percentage of individuals in the upper and
lower 5-cm extremes (panels c,d and e,f respectively) in Experiment 3. Left panels give
results for Bosmina longirostris and right panels give results for Daphnia pulicaria.
Values reported are treatment means (i SE). Treatments included: control water or water
with kairomones from the small invasive predator, Cercopagis (CP), or the large, native
predator, Perca (PF). Asterisks indicate treatments that were significantly different (p <
0.05) from the control.

43

depth of B. longirostris was significantly deeper for the C ercopagis treatment than the
control treatment and did not differ between Perca and control treatments (F 2,27 = 4.71, p
= 0.02; Figure 2.3a). The percentage of B. longirostris at lower extremes of the column
was significantly greater for the Perca treatment than the control and did not differ
between Cercopagis and control treatments (Figure 2.3a), while the percentage of B.
longirostris at the upper extremes did not significantly vary among treatments (all ps >

0.5).

DISCUSSION

Our results indicate that Lake Michigan zooplankton generally lack naiveté
towards invasive predators. As shown for Bythotrephes in this study and in a previous
study (Pangle and Peacor 2007), three species of zooplankton all responded adaptively to
the invasive predator C ercopagis, despite having only six years experience with the
predator. In the first and third experiments, Cercopagis induced downward migration in
the relatively vulnerable prey species, Bosmina, but not in less vulnerable prey species,
D. mendotae and D. pulicaria. Although this study is the first of our knowledge to
experimentally evaluate the response of zooplankton to Cercopagis, the differential
response to C ercopagis corresponds with field observations in Lake Ontario, in which
small zooplankton species exhibit deeper distributions and less overlap with Cercopagis
relative to large zooplankton species (Benoit et al. 2002).

Our results also suggest that migration induced by the invasive predators is a
more general trait that is also induced by native predators. In the first experiment, both

Bythotrephes and the functionally similar, but phylogenetically disparate native predator

44

Perca induced downward migration in D. mendotae. In contrast, D. mendotae exhibited
no migratory response to the native predators Leptodora and Polyphemus, which are
much more phylogenetically but less functionally similar to Bythotrephes relative to
Perca. Further, D. mendotae migrated upwards in response to kairomones of Mysis, a
predator that is functionally and phylogenetically disparate to Bythotrephes relative to the
other native predators. Although not predicted a priori, upward migration seems an
appropriate response to the deep dwelling Mysis and is consistent with field observations
of zooplankton avoidance of Mysis in Lake Michigan (Schulze and Brooks 1987).

The behavioral responses we observed in zooplankton therefore challenge the
notion that native prey in aquatic systems are generally naive to invasive predators. The
basis for this notion is that aquatic prey tend to lack evolutionary experience and thus
possibly appropriate traits to respond to predators outside of their system. However,
characteristics of zooplankton may allow them to perceive and respond adaptively to
novel risk. First, zooplankton can assess the level of risk via kairomones, as multiple
studies have shown that zooplankton respond more strongly to kairomones of predators
that had recently consumed conspecifics, than to predators that were unfed or fed other
zooplankton species (Stabell et a1. 2003; Laforsch et al. 2006). Thus, consumption of
conspecifics by invasive predator may be sufficient to alert zooplankton of their presence.
Second, zooplankton tend to use qualitatively-generalized antipredator traits that are
adaptive for multiple predators. For example, growing larger helmets and tail spines is
adaptive for Daphnia facing three functionally similar invertebrate predators (Laforsch
and Tollrian 2004). In our study, downward migration is a trait that likely evolved due

fish predation, but may have been co-opted for Cercopagis and Bythotrephes. Although

45

we have focused here on zooplankton, these characteristics, and thus mechanisms that
reduce naiveté, are also present in other aquatic prey that use kairomones to perceive risk
and have qualitatively-generalized responses to multiple predators (e. g., protozoa
[Kuhlmann et al. 1999], rotifers [Stemberger and Gilbert 1987], and amphibians [Relyea
2001]).

However, it remains unclear how zooplankton came to associate Bythotrephes and
C ercopagis with the appropriate predator functional type. For example, why did
zooplankton not migrate upwards in response to Bythotrephes, as they did in the presence
of Mysis? One explanation is that downward migration rapidly evolved in response to
selection pressures of the invasive predators. The rapid evolution hypothesis is supported
by changes observed in prey populations in Lake Michigan following the introduction of
invasive predators. The density of D. mendotae was extraordinarily low during the year
following the introduction of Bythotrephes, but then rebounded to pre-invasion densities
during subsequent years (Lehman and Caceres 1993). These population changes may
represent a bottleneck in which strong selection occurred for Daphnia genotypes that
migrated downwards in Bythotrephes presence. There is growing evidence that
evolutionary processes generally contribute to the ecology of invasions over relatively
short time scales (reviewed in Mooney and Cleland 2001; Strauss et a1. 2006).
Zooplankton have characteristics that are particularly conducive to rapid evolution, such
as short generations and cyclo-parthenogenic reproduction, and have been shown to
evolve rapidly in response to changes in their environment over periods as short as ten
years (Hairston et al. 1999; Cousyn et al. 2001; Kerfoot and Weider 2004). Although we

are unable to test this explanation given the methods of our study, future research may be

46

able to do so by hatching resting eggs of zooplankton collected from Lake Michigan that
were deposited before and after the introduction of Bythotrephes and C ercopagis and test
whether the response of zooplankton changed over this time period. Such research could
not only provide insights on the role rapid evolution plays in the prey nai'veté, but also
identify a mechanism that leads to the change of interactions between native and invasive
species often observed over time (Strayer et al. 2006).

Identifying mechanisms underlying the detrimental impact of invasive species is
one of the most important challenges facing biologists today (Mack et al. 2000; Byers et
al. 2002; Hochberg and Gotelli 2005).Na‘1'vete’ is thought to exacerbate this impact,
particularly in lakes (Cox and Lima 2006); however, results of this study indicate that
some lake species are not na'ive, thus changing the interactions predicted between invader
and prey. Although migration is expected to reduce direct consumption of zooplankton
by Bythotrephes and Cercopagis, migration may lower birth rate of zooplankton due to
inhabiting regions of lower temperature (Chapters 3, 5), referred to as a non-consumptive
effect (Abrams 2007). These non-consumptive effects may lead to large reductions in
prey population growth rate (Chapter 5), and change how zooplankton interact with
competitors and other predators due to the shift in zooplankton habitat use (Chapter 6), a
trait-mediated indirect effect (Abrams 2007). Thus, although prey are not naive,
understanding how they are responding to invasive predators may still paramount to

understanding the ecology of invaded systems.

47

CHAPTER 3

TEMPERATURE GRADIENTS, BUT NOT FOOD RESOURCE GRADIENTS,
AFFECT GROWTH RATE OF DAPIHVIA MENDOTAE IN LAKE MICHIGAN

ABSTRACT
We evaluated the effects of water temperature and food resource gradients on the growth
rate of zooplankton undergoing vertical migration in Lake Michigan. In two laboratory
experiments, juvenile Daphnia mendotae, a native herbivorous cladoceran, were
incubated for five days at water temperatures associated with the epilimnion and deep
chlorophyll maxima (DCM) of Lake Michigan and fed food resources collected directly
from these regions. Juvenile growth rate was strongly dependent on water temperature, as
Daphnia incubated at the epilirnnetic temperature (21°C) exhibited a 42 % higher growth
rate than those at the DCM temperature (8°C). Juvenile growth rate of Daphnia that
alternated between the two temperatures every 12 hours (10.8 % day'l) was similar to the
arithmetic average growth rate of the two extreme water temperature treatments (11.0 %
day"), suggesting fluctuating temperatures itself does not substantially influence Daphnia
growth. Contrary to water temperature, the depth at which food resources were obtained
had no effect on juvenile growth rate of Daphnia, nor was there a significant interaction
between food resource and water temperature effects. The results of our study show a
dominant influence of variation in water temperature relative to variation in food
resources on the growth rate of migrating Lake Michigan Daphnia and demonstrate the
need to account for the different water temperatures Daphnia encounter over the course

of a day when estimating their production. These findings may extend to other migrating

48

zooplankton species and to other Laurentian Great Lakes, though gradients other than

water temperature may also need to be considered (e.g., oxygen gradient in Lake Erie).

INTRODUCTION

The production of zooplankton is important to Great Lakes ecosystems, as
zooplankton serve as a critical food resource to many fish species (Miller et al. 1990;
O'Gorman et al. 1997; Fulford et a1. 2006) and may also influence density and
composition of phytoplankton (Makarewicz et al. 1998), which may in turn affect water
clarity of the lakes (Scavia et al. 1986). Making accurate predictions of zooplankton
production relies on knowledge of the environment they inhabit (Shuter and Ing 1997;
Kuns and Sprules 2000); however, predictions may be complicated by the migratory
behavior of zooplankton. Many species of zooplankton in the Laurentian Great Lakes
undergo diel vertical migration (Wells 1960; Schulze and Brooks 1987; Barbiero et al.
2000; Chapter 5), a behavior that is general to freshwater and marine systems (DeMeester
et al. 1998). Although the ultimate causes of migration by zooplankton may vary among
lakes (Leech et a1. 2005), migration may occur in the Great Lakes as means for
zooplankton to avoid predation risk (Schulze and Brooks 1987; Chapter 5). A
consequence of migration is that the environment zooplankton experience can greatly
change over the course of a day, particularly when migration occurs during periods of the
year when the lakes are vertically stratified. Changing environments could influence the
production of zooplankton, but exactly how is unclear; there are multiple environmental
gradients associated with the water column, and to our knowledge, the effects of these

gradients on zooplankton migrating in the Great Lakes have yet to be evaluated.

49

Water temperature and food resources are two vertical gradients that have been
shown to influence zooplankton growth and reproduction in other systems (Williamson et
al. 1996; Winder et al. 2003; Kessler and Lampert. 2004). During stratified periods,
zooplankton migrating in lakes will generally encounter a water temperature gradient,
and the amount of time zooplankton spend at each temperature stratum may strongly
influence their metabolic and reproductive rates (Bottrell 1975; Orcutt and Porter 1983;
Stich and Lampert 1984). Migrating zooplankton may also encounter variation in food
resources, particularly in lakes that exhibit deep chlorophyll maxima (DCM). The DCM
is a peak in phytoplankton biomass that is observed in lakes, including the Great Lakes
(Barbiero and Tuchman 2001), at the base of the thermocline, and zooplankton fed food
resources from the DCM have been shown to sometimes exhibit greater growth and
reproduction than zooplankton fed food resources from shallow regions of the water
column (Williamson et a1. 1996; Winder et al. 2003; Park et al. 2004; Kessler and
Lampert 2004). The relative importance of temperature and food resource gradients to
zooplankton growth has been shown to vary considerably among different lakes, with
temperature being relatively more important in some lakes (Williamson et al. 1996; Park
et al. 2004) and food resources being relatively more important in others (Cole et a1.
2002; Winder et al. 2003). This variation among lakes may be influenced by differences
in the severity of the water temperature gradient (Winder et al. 2003) and zooplankton's
utilization of food resources other than phytoplankton, such as ciliates, bacteria, and
detritus (Rothhaupt 1991; Williamson et al. 1997).

In this study, we evaluated the effects and relative importance of temperature and

food resource gradients found in Lake Michigan on the growth of the zooplankton

50

species, Daphnia mendotae. Daphnia mendotae is a dominant zooplankton species during
the summer in Lake Michigan (Barbiero and Tuchman 2001b) and is known to undergo
diel vertical migration in the lake, aggregating near the DCM during the day and in the
epilimnion during the night (Chapter 5). Using laboratory experiments, we mimicked the
variation in environments experienced by D. mendotae, by incubating them at water
temperatures of and feeding them food resources collected from either the DCM or the
epilimnion of Lake Michigan. Our results demonstrate the relative effects of water
temperature, food resources, and their possible interactions on D. mendotae in Lake

Michigan and have implications to estimating zooplankton production in the Great Lakes.

METHODS

We conducted two experiments during the summer 2005 in which we evaluated
juvenile Daphnia growth rate under different water temperature and food resource
treatments. Our treatments were designed to 1) tease apart the effects of water
temperature and food resource gradients associated with vertical migration by varying
one of the factors while holding the other constant; 2) identify possible interactions
between the effects of water temperature and food resources; and 3) determine if
fluctuating environments themselves affect growth rate. We chose to use juvenile growth
rate as a response variable, because it is strongly correlated to the intrinsic rate of
population growth (Lampert and Trubetskova 1996). To ensure that Daphnia did not
mature over the course of the experiment, we limited the duration of each experiment to 5

days. For both experiments, we obtained food resources and Daphnia from Lake

51

Michigan in an identical fashion, which we describe first before presenting the specific
details of each experiment.

Food resources fed to Daphnia during the experiments was seston, including
phytoplankton, ciliates, bacteria, and detritus, collected at ambient densities from the
DCM and epilimnion of Lake Michigan. We were unable to make daily seston collections
from Lake Michigan due to logistical limitations. Instead, seston was collected once two
days prior to the start of each experiment and maintained over the duration of the
experiment in light and temperature conditions that were similar to the depths seston was
collected from following techniques used by Fahnenstiel et al. (2000) and Lohrenz et al.
(2004). Collections for the first and second experiments occurred on 22 August and 17
September 2005, respectively, at a site approximately 10 km west of Muskegon,
Michigan. Prior to collection, a vertical profiler (Sea-Bird Electronics, Inc., Bellevue,
WA) equipped with an in situ chlorophyll fluorometer (Seapoint Sensors, Inc., Exeter,
NH) was deployed to identify thermal structure of the water column and the location of
the DCM. Vertical profiles of water temperature and chlorophyll concentration on the
collection dates are presented in Figure 3.1. Seston was collected by pumping lake water
from the center of the epilimnion and the DCM (Figure 3.1). Pumped lake water was
passed through a 300-um sieve and into separate 25—L transparent, polyCarbonate
carboys. The sieve retained planktivores (e.g., Bythotrephes longimanus, Leptoa'ora
kindtii), while allowing the seston and herbivorous zooplankton to remain in the collected
lake water at ambient densities. The carboys were immediately placed in dark coolers and
transported to the National Oceanic and Atmospheric Administration Lake Michigan

Field Station in Muskegon, MI. At the field station, carboys were maintained for seven

52

Water temperature (°C)
4 812162024 4 812162024

 

 

 

 

 

O l l l l
10 - -
E
5 201 '4
a.
a)
O
l
30 a l -
l
l —— Water temperature
‘ -—— Chlorophyll
40 - -

 

 

 

 

I If T

0.03 0.06 0.09 0.12 0.03 0.06 0.09 0.12
In situ chlorophyll a (fluorescence units)

Figure 3.1. Water temperature (solid line) and in situ chlorophyll fluorescence (dashed
line) in Lake Michigan during water collection for the first and second experiments
(panels a and b, respectively). Arrows indicate the depths at which food resources were
collected from for each experiment.

53

days (two days prior and five days after the start of the experiment) in translucent, water-
filled environmental chambers that mimicked light and temperature conditions observed
in the epilirnnetic and DCM regions of the water column, respectively. The
environmental chambers were exposed to full sunlight and were lined with theatrical gels
to manipulate the spectra of sunlight entering the chambers. Water temperature inside the
chambers was regulated using an external chiller. To avoid the settling of seston, the
carboys were gently rotated three times end-over-end every 12 hours. Our maintenance of
lake water, including the use of these specific chambers, has been shown to maintain
ambient densities of seston for at least 10 days (G. L. Fahnenstiel, personal
communication, 2004).

Daphnia used in the experiments were neonates born from adults collected from
Lake Michigan. Adult Daphnia were collected at the same time as lake water using a 0.5
diameter 64-pin zooplankton net with a closed bucket towed vertically through the entire
water column. From the netted zooplankton, 200 egg-bearing Daphnia were transferred
into a 2-L container 13 hours prior to each experiment, and neonates born from these
Daphnia were collected one hour prior to each experiment. An additional 25 neonates
were collected for an estimate of initial size.

At the start of each experiment, four Daphnia neonates were transferred into 300-
ml replicate bottles filled with either epilirnnetic or DCM lake water. Bottles were housed
in incubators for the duration of the 5-day experiments and rotated at 0.25 rpm using
plankton wheels. Incubators maintained the bottles at different experimental temperatures

and regulated photoperiod to 14 h light : 10 hr dark.

54

In our first experiment, we used a factorial design, with six treatments in total: we
used two food resources (epilirnnetic and DCM seston) under three thermal regimes
(Table 3.1). The first two thermal regime treatments consisted of exposing Daphnia to
either 21 or 8°C, which are typical water temperatures found in the epilirnnetic (21°C) or
DCM (8°C) region of the water column. The third treatment consisted of exposing
Daphnia to alternating temperatures of 21°C and 8°C, switching at 12-hour intervals.
Based on the approximate linear relationship between water temperature and grth of
Daphnia under non-saturating food resource conditions (Hall l964),we expected that
growth rate in the alternating temperature treatment would be approximately the
arithmetic average of growth rates of the high and low temperature treatments; a grth
rate lower than this average would indicate a growth reduction due to fluctuating
temperatures (Reichwaldt et al. 2005). The factorial design also allowed us to evaluate
the interacting effects that may occur between food resource conditions and temperature
on Daphnia growth. For example, Gielbelhausen and Lampert (2001) showed that
temperature had a greater effect on Daphnia magna grth in high relative to low
concentrations of food resources. Each treatment in the first experiment had four replicate
bottles (total, 24 bottles), and Daphnia in all treatments were transferred into a new bottle
filled with fresh treatment water every 24 hours. Transfers were done to ensure I) that
Daphnia did not consume more than 25% of the available food resources, based on
known clearance rates, and 2) that bottles did not foul.

In our second experiment, we tested the effect of simultaneous changes in water
temperature and food resources that Daphnia would experience when migrating in Lake

Michigan. Specifically, we included one treatment in which we alternated epilimnion and

55

Table 3.1. Temperature and food resource treatments used in the first and second

 

 

experiments.
Treatment Temperature (°C) Food resource
Experiment 1 1 21 Epilimnetic
2 21 DMC
3 8 Epilimnetic
4 8 DMC
5 alternate 8 and 21 Epilimnetic
6 alternate 8 and 21 DMC
Experiment 2 1 alternate 8 and 21 Epilimnetic at 21°C and DCM at 8°C
2 alternate 8 and 21 Epilimnetic
3 alternate 8 and 21 DCM
4 21 Epilimnetic

 

DCM = deep chlorophyll maxima

56

DCM water temperature and food resources every 12 hours and two treatments in which
we alternated water temperature but not food resources (Table 3.1). Comparing the first
treatment with the latter two provided a means to determine whether Daphnia benefited
from the combination of food resources and water temperature experienced during
migration. We also included a fourth treatment in which Daphnia were maintained at a
constant 21°C and fed epilirnnetic food resources. The difi‘erence between this treatment
and the first treatment provided an estimate of the grth cost incurred by migrating
Daphnia relative to remaining the epilimnion. In addition, the last three treatments were
similar to treatments in the first experiment, thus providing a means to reaffirm results of
the first experiment with food resources taken on a different date. Each treatment in the
second experiment had four replicate bottles (total, 16 bottles), and Daphnia in all
treatments were transferred into a new bottle filled with fresh treatment water every 12
hours.

For both experiments, juvenile growth rate (g, day") was calculated as:

ln(lf) -— ln(ll.)

(3-1) g= d

 

where lfand I,- were the final and initial standard length and d was number of

experimental days. The initial length of Daphnia (mean i standard deviation) in the first

and second experiments was 552 i 22 and 566 i- 64 pm, respectively. Mean survival rate
of Daphnia over the course of the experiments varied among treatments, ranging from 89
to 94 % in both the first and second experiments. To ensure that juvenile growth rate was
not confounded by survival, we conducted two tests. First, we compared survival rate

among treatments using ANOVA. Second, we evaluated the effect of survival on growth

57

rate among replicates of all treatments using linear regression. The tests showed that
survival did not differ among treatments and did not significantly affect juvenile growth
rate.

We analyzed the effects of water temperature and food resource treatments on
juvenile growth rates using general linear models. For the first experiment, we evaluated
the effect of water temperature and food resources and their interaction using a two-way
ANOVA. For the second experiment, we determined if treatments varied significantly
using a one-way ANOVA. If a significant difference was found, we used a Tukey's HSD

test to separate treatment differences.

RESULTS

In the first experiment, the juvenile growth rate of Daphnia significantly varied
over the range of temperatures associated with the water column in Lake Michigan (F 2,20
= 37.37, p < 0.01). Daphnia incubated at the epilirnnetic water temperature (21°C) grew
at a 42 % greater rate than those at the DCM water temperature (8°C) and a 29 % greater
rate than those that alternated between these water temperatures (Figure 3.2). The average
juvenile growth rate of Daphnia switched between the two temperatures (0.108 day'l) fell
close to the arithmetic average growth rate of the two extreme water temperature
treatments (0.110 day'l). Contrary to water temperature, the juvenile growth rate of
Daphnia was not significantly affected by the location that their food resources were
collected from (F 1,20 = 0.66, p = 0.43). Daphnia grew at a similar rate in epilirnnetic and

DCM water under each temperature treatment, as further indicated by the non-significant

58

0.18 ‘ l:l Epilimnion
C: DCM

0.15 ~ iL

0.12 -
0.09 —

0.06 a

Juvenile growth rate (d'1)
tl-l
d—l

0.03 -

 

 

 

 

 

 

 

 

 

0.00 . l ,
8°C 8 to 21°C 21°C

Water temperature treatments

 

Figure 3.2. Growth rates of Daphnia in Experiment 1. Juvenile Daphnia experienced
three temperature treatments (8°C, alternating 8 and 21°C, and 21°C) and two food
resource treatments (water from the epilimnion and the deep chlorophyll maxima
[DCM]). Values reported are treatment means (i SE).

59

interaction between food resource and water temperature effects (F 2,13 = 1.26, p = 0.31).
The interaction term was subsequently removed from the ANOVA.

In the second experiment, the juvenile growth rate of Daphnia indicated that
simultaneous changes in water temperature and food resources did not have a substantial
effect on Daphnia relative to only changes in water temperature. Although juvenile
growth rate varied significantly among treatments (F 3,12 = 16.51, p < 0.01), growth of
Daphnia that experienced alternating epilirnnetic and DCM water temperatures and food
resources were similar to those that experienced alternating epilirnnetic and DCM water
temperatures, but fed either epilimnetic or DCM food resources (Figure 3.3). Daphnia
that experienced constant epilirnnetic water temperature and food resources grew at 34 %
greater rate than those in the three alternative temperature treatments on average (Figure

3.3).

DISCUSSION

Daphnia migrating in Lake Michigan experience gradients in both water
temperature and food resources; however, our results indicate that changes in water
temperature has a much greater effect on Daphnia than changes in food resources. In
both experiments, increased water temperature had positive effects on Daphnia growth
rate, most likely caused by a heightening of metabolic rates (Orcutt and Porter 1983;Stich
and Lampert 1984). Contrary to water temperature, food resources taken from the deep
chlorophyll maxima did not improve Daphnia growth rate relative to epilirnnetic food

resources. There was also no interactive effect between water temperature and food

60

0.18 -

 

 

 

 

 

 

 

 

y
A .L
.> 0.15 —
g z z z
3 0.12 ~ I
E
g 0.09 _
9
U)
9 0.06 —
'E
0.)
:>: 0.03 -
_,
0.00 - . . .
DCM to Epi. Epi. DCM Epi.

8to 21°C 8to 21°C 8to 21°C 21°C

Water source and temperature treatments

Figure 3.3. Growth rates of juvenile Daphnia in Experiment 2. Daphnia experienced one
of four treatments including: deep chlorophyll maxima (DCM) water at 8°C for 12 h and
epilirnnetic water at 21°C for 12 h; epilirnnetic water at 21 °C; epilirnnetic water at 21°C
for 12 h and at 8°C for 12 h; and DCM water at 21°C for 12 h and at 8°C for 12 h. Values
reported are treatment means (:h SE). Means with different lowercase letters were
significantly different as determined by AN OVA and Tukey’s HSD tests (p < 0.05).

61

resources. A previous study that had found such interactions concluded that they occur
only when food resources go from a limiting to non-limiting state (Gielbelhausen and
Lampert 2001). Thus, the lack of interaction may have occurred because food resources
were limiting for D. mendotae in both the DCM and the epilimnion. In addition, although
food resources for the experiments were collected on different dates (August and
September), the consistent lack of a food resource effect suggests that our findings are
general to the late summer and early fall months during which Daphnia are most
abundant (Pangle, unpublished data). Because the vertical gradient in phytoplankton
biomass is generally most severe during this period (F ahnenstiel and Scavia 1987), the
food resource gradient may be even less important during other months of the year.
While ours is the first study to evaluate effects of migration on zooplankton in the
Great Lakes, other studies have explored these effects in other lake systems. In a lake that
had a similar temperature gradient and a more severe food resources gradient than Lake
Michigan, Park et al. (2004) showed Daphnia rosea fed DCM seston grew faster than
those fed epilimnetic seston; however, consistent with our results, this food resource
effect was relatively minor compared to corresponding water temperature effects. Similar
results have been shown for both Daphnia pulicaria and a copepod, Diaptomus
oregonensis (Williamson et al. 1997), indicating that the findings of our study may also
be applicable to other non-Daphnia zooplankton in Lake Michigan. Contrary to our
results, Winder et al. (2003) found that the food resources gradient in a Switzerland lake
had a greater influence on the growth of Daphnia galeata than water temperature
gradient. However, their water temperature gradient was much less severe than in Lake

Michigan, with only a change of 6°C. Further, food resource gradients may vary among

62

systems in ways that are difficult to quantify, particularly because Daphnia consume food
other than phytoplankton, including ciliates, bacteria, and detritus, with maxima that may
not correspond with the DCM (Rothhaupt 1991; Williamson et al. 1997).

The results of our study have multiple implications on estimating zooplankton
production in Lake Michigan. First, our study experimentally validates the assumption
taken by many studies that production of zooplankton is a function of water temperature
(e. g. Shuter and Ing 1997). Second, our study indicates that production estimates should
consider the possible variation in water temperature experienced daily by zooplankton.
Doing this may require depth-stratified sampling of zooplankton, as described in a
previous study (Kuns and Sprules 2000), carried out several times over 24-hours periods.
Although our study focused on Lake Michigan, these implications may have direct
applications to the other Laurentian Great Lakes, as these lakes can exhibit similar
temperature and food resource gradients to Lake Michigan (Barbiero and Tuchman
2001a). However, other gradients may also be in important in the other lakes. In
particular, oxygen gradients that can develop in Lake Erie during the summer associated
with the "Dead Zone" (Rosa and Burns 1987) has the potential to strongly influence
zooplankton production (Weider and Lampert 1985; Wright and Shapiro 1990).

Our study also shows how planktivores can affect zooplankton in Lake Michigan
through means other than direct predation. Vertical migration by Daphnia mendotae
functions as means to avoid predation risk and varies according to densities of shallow-
dwelling planktivores such as Bythotrephes (Chapter 5). Thus, the cost of migration, as
demonstrated in our experiments, can be considered a non-consumptive effect of

planktivores on Daphnia populations. Non-consumptive effects have been shown to

63

contribute substantially to the net effect of predators and influence predator-prey
dynamics in qualitatively different ways than direct consumption of prey (Ives and
Dobson). Incorporating these non-consumptive effects is therefore necessary to
understanding dynamics Daphnia and other migrating zooplankton in the context of the

Lake Michigan food web.

64

CHAPTER 4

LIGHT-DEPENDENT PREDATION BY THE INVERTEBRATE PLANKTIVORE,
B YT HOT REPHES LONGIMAN US

ABSTRACT
We evaluated predation by the invasive invertebrate planktivore, Bythotrephes
longimanus, on a Lake Michigan prey assemblage as a function of light intensity.
Daphnia mendotae was the only prey type significantly reduced and light intensity
strongly affected this reduction. Specifically, Bythotrephes consumption of D. mendotae
was not detected under low light intensity (< 1 mo] m2 s"), but increased with greater
light intensity, and leveled off under high light intensity (> 100 mol m'2 s"), at which
point Bythotrephes ingestion rate was 2.1 Daphnia/h. These results indicate that
Bythotrephes predation is more sensitive to light than previously thought, a discrepancy
that can be explained after considering the ability of D. mendotae to detect Bythotrephes'
hydromechanic disturbance. The observed effect of light intensity on Bythotrephes
predation is more like that of planktivorous fish than that of other previously-studied
invertebrate planktivores. Our findings elucidate the role Bythotrephes plays in the food
web and provide a novel explanation for its tendency to invade lakes of high water
clarity. The importance of light-dependent predation found here may extend to other

visually oriented predatory cladocerans.

INTRODUCTION
Light can strongly influence the composition and dynamics of communities in

aquatic systems (Interlandi and Kilharn 2001; Wissel et al. 2003; Semes and Aksnes

65

2006). One mechanism underlying this effect relates to the role light plays in predator-
prey interactions (Aksnes 2007). For example, some planktivores have consumption rates
that are order-of-magnitudes greater in lit environments than in the dark because they rely
on visual cues to detect prey (Ali 1959). Such predation pressure can drive visually-
conspicuous prey to extirpation (Brooks and Dodson 1965) and shift habitat use of
remaining prey (Clark and Levy 1988; DeMeester et al. 1998), both having potentially
strong indirect effects on aquatic communities (Carpenter and Kitchell 1993).

Since planktivores can play a dominant role in aquatic communities, the potential
consequences of the invasive predatory cladoceran, Bythotrephes longimanus, in the
Laurentian Great lakes and surrounding inland lakes has received much attention
(Lehman and Caceres 1993; Vanderploeg et al. 1993; Boudreau and Yan 2003; Barbiero
and Tuchman 2004; Strecker et al. 2006). Since its arrival in the Great Lakes two decades
ago, Bythotrephes has been implicated with the extirpation and reduction in density of
several zooplankton species (Lehman and Caceres 1993; Yan et al. 2002). This disruption
to pelagic food web can have indirect consequences on other organisms, including fish
(Hoffman et al. 2001) and rotifers (Hovius et al. 2007). Understanding the long-term
effects of Bythotrephes will require continued monitoring (N eilson et al. 2003) and a
better understanding of the mechanisms governing its effect (Schulz and Yurista 1999,
Boudreau and Yan 2003).

Muirhead and Sprules (2003) showed that Bythotrephes’ ability to detect prey
increased with increased light intensity using experiments that measured the response of
tethered Bythotrephes to prey. This type of prey detection is different from that associated

with more commonly studied invertebrate planktivores, such as Chaoborus spp. and

66

predatory copepods, which depend primarily on hydromechanical cues for tactile
detection of prey. In confirmation of the importance of light, Bythotrephes generally
inhabit the most lit portions (i.e., epi- and metalimnion) of the water column (Evans
1988; Lehman and Caceres 1993). Further, field surveys (Lehman and Caceres 1993,
Chapter 5) and laboratory experiments (Chapter 1) indicate that Bythotrephes induce a
strong downward migration in prey during the day, but not at night. Such behavioral
responses are presumably an adaptive response to reduce predation risk associated with
light intensity (reviewed in DeMeester et al. 1998).

Despite its importance to accurately quantifying and modeling Bythotrephes
predation rate as a fimction of time and among systems, the functional relationship
between Bythotrephes predation and light intensity is still not well understood. To our
knowledge, there has not been a study that has experimentally measured predation rate of
Bythotrephes at different light intensities. This gap in knowledge is also true for
predatory cladocera in general, despite their dominance in many lakes (Manca et al.
2000; Pichlova and Brand] 2003; Halversen et al. 2004; McNaught et a1. 2004; Ojaveer et
al. 2004; Wojtal et al. 2004). Further, when quantifying Bythotrephes predation rate using
encounter models (Muirhead and Sprules 2003), Bythotrephes is predicted to consume
prey even in the absence of light; however, we found that Bythotrephes predation rates
were below detection limits in an experiment using low light intensity (< 0.1 umol m'2 s'
l; K. L. Pangle, unpublished data). The goal of this study was to evaluate the effect of
light intensity on Bythotrephes predation using experiments designed to encompass the

variation of light intensity and prey observed in nature. Our findings will improve

67

predictions of Bythotrephes influence on pelagic communities across its current and

expanding geographical range.

METHODS

We conducted two separate predation experiments in which Bythotrephes were
incubated with prey at different light intensities, one occurring on 1 September 2005 and
the other on 13 October 2005. Each experiment had three light intensity treatments,
which we labeled high, mid and low (Table 4.1). The second experiment explored a light
intensity gradient lower than the first gradient; the lowest light intensity was 30 times
lower than the lowest light intensity in the first experiment, with the middle and highest
intensities approximating those from lowest and middle light intensities from the first
experiment (Table 4.1).

The prey consisted of a natural assemblage of zooplankton to make our results
most applicable to predation occurring in natural systems (Colton 1987). For both
experiments, 4-L glass bottles containing Bythotrephes and prey were submerged in an
incubation chamber, and predation rate was estimated by comparing the final density of
zooplankton in bottles that contained Bythotrephes to that in control bottles lacking
Bythotrephes. To manipulate light intensity among treatments, bottles were wrapped with
varying layers of neutral density vinyl screening to reduce the fi'action of light entering
the bottle during the experiment. We designed the experiments so that the loss of
zooplankton due to Bythotrephes predation would be kept low enough to avoid severe

prey depletion, and high enough to still be detectable statistically (V anderploeg et al.

68

Table 4.1. Summary of the different Bythotrephes densities and light intensities used for
each treatment in Experiments 1 and 2.

 

 

Light Bythotrephes Light intensity (umol m'2 3")
treatments per bottle minimum mean maximum
Experiment 1 High 1 55 170 397
Mid 1 8 25 60
Low 3 1 3 8
Control 0 55 170 397
Experiment 2 High 1 8 22 71
Mid 3 1 3 11
Low 3 0.02 0.06 0.21
Control 0 8 22 71

 

69

1982; Bamstedt et al. 2000). Zooplankton densities were also chosen to be within a range
observed in the field (K. L. Pangle, unpublished data).

Our incubator chamber allowed for 36 bottles per experiment. Consequently, we
decided to only have one control treatment, using the high light intensity for each
experiment, thus allowing for more replicates of each treatment. The high light intensity
treatment was chosen because it is in this treatment that potential of prey loss is highest in
Bythotrephes absence, because predation rate is likely greatest from other predators in the
community sampled (e.g., Epischura lacustris, Polyphemus pediculus, and Leptodora
kindtii), which may also display light-dependent predation (Young and Taylor 1987).
While we expected minor prey loss in the absence of Bythotrephes, we believe our use of
a high light intensity control would, if anything, lead to underestimates of the effects of
light intensity on Bythotrephes predation rates, and thus is conservative in the context of
this experiment.

On the day of each experiment, live Bythotrephes and zooplankton prey were
obtained from Lake Michigan at a site 5 km west of Muskegon, Michigan (43° 11'29"N,
86° 25'92"W, bottom depth = 45 m). Prior to the initiation of each experiment,
approximately equal amounts of zooplankton were added to each Bythotrephes treatment
bottle (27 bottles), each control bottle (9 bottles). An additional 18 samples were also
taken and used to assess initial zooplankton densities. At the start of each experiment,

Bythotrephes were added into Bythotrephes treatment bottles (27 bottles total), while

70

control treatment bottles received no Bythotrephes. We anticipated lower predation rates
as light intensity decreased, and therefore added more Bythotrephes to treatments with
lower light (Table 4.1).

Bythotrephes were collected by towing a l-m diameter, 500-um-mesh plankton
net equipped with a 4-L collection bottle from 25 meters depth to the surface. Captured
Bythotrephes were immediately transferred into 15-ml vials (1 individual per vial) that
were filled with lake water that had been passed through a 64-micron filter to remove all
zooplankton prey. To reduce variability among Bythotrephes, we separated individuals by
instar and used individuals of the 2nd instar in the experiments (mean individual weight
in both experiments was 0.3 mg DW). The experiment started when Bythotrephes were
poured from the vials into the experimental bottles.

Zooplankton prey were collected using a 0.5 m diameter, 64-um-mesh plankton
net equipped with a l-L collection bottle at the cod end towed over the same area.
Captured zooplankton prey were gently poured into a 30-L aquarium, which was filled
with filtered lake water. Cetyl alcohol was added to the aquarium to prevent zooplankton
from adhering to the surface of the water (Desmarais 1997). Zooplankton prey were
added to the experimental bottles by taking lO-ml subsamples of water from the center of
the aquarium and releasing subsamples into the bottles. We took subsamples by lowering
an 8-mm diameter cylinder into the water until it touched the bottom of the aquarium at
which point we capped the top of the cylinder to retain the water and organisms inside.
This subsampling approach allowed us to capture organisms regardless of their vertical
position in the aquarium and also did not require producing a suction field that could

potentially be sensed and avoided by some zooplankton. Subsamples were transferred to

71

bottles according to a randomized-block design, a technique that minimized the variation
in zooplankton density between bottles, and one subsample was added to each bottle
(both the 36, 4-L experimental bottles and the 18, 0.5-L sample bottles was used to
estimate initial zooplankton densities) sequentially, until each bottle received 16
subsamples total. Initial zooplankton densities samples were preserved with buffered
sugar-formalin solution.

The incubation chamber in which experiments were carried out was 2,000-L in
size and maintained an environment similar to that observed in Lake Michigan. The
incubation chamber was illuminated with natural sun. Light entering the chamber passed
through theatrical gels (Peacock Blue Roscolux, Rosco Laboratories Inc., Stamford,
Connecticut, USA) that mimicked the pelagic light spectra and neutral density vinyl
screening that reduced light intensity (Lohrenz et al. 2004). Ambient light intensity was
measured using a 4-pi sensor over the course of the experiments, and the number of
neutral vinyl layers was changed as sunlight intensity changed. In comparison with light
intensity observed in Lake Michigan's water column, 25 and 170 mol m'2 3" would be
typical at the thermocline and at the center of epilimnion, respectively, on a summer day,
whereas 0.06 umol m'2 5'1 would be typical at the surface on a moonlit night (K. L.
Pangle, unpublished data). Water temperature inside the incubation chamber was kept at
a level typical of the Lake Michigan metalimnion (16°C) using an external water chiller.

Bottles were attached to a motorized cage located inside the incubation chamber,
which rotated the bottles end-over-end during the experiment at a constant rate
(approximately 0.5 rpm) to reduce aggregation of zooplankton inside the bottles. Both

experiments ran for 7 hours, from 1300 to 2000 EDT. At the end of each experiment,

72

bottles were removed from the incubation chamber, and organisms were immediately
narcotized by adding Alka-Seltzer® tablets to each bottle. Organisms were siphoned
from the bottles onto a 64-urn filter and then transferred into 0.2 L bottles with buffered
sugar-formalin solution. Later, all zooplankton prey and Bythotrephes were identified and
counted using a microscope and their body lengths were measured using a drawing tube
and digitizer (Roff and Hopcroft 1986). Prey were grouped into one of 7 prey types
including: Daphnia mendotae, Bosmina Iongirostris, copepod nauplii, juvenile
Diaptomus spp., adult Diaptomus spp., juvenile Diacyclops thomasi, and adult D.

thomasi.

Statistical analyses and predation rate calculations

Prior to analyses of light effect, we first tested assumptions of normality and
homogeneity of variance using the Wilks-Shapiro Test and the Levene's Test,
respectively. In addition, we evaluated whether prey density was substantially different in
the control bottles at the end of the experiment than it was at the start of the experiment
due to predation by other predators in the community sampled. This was done by
comparing the average initial prey density based on the preserved initial samples with the
average prey density in the control bottles at the end of the experiment. While the
predatory copepod, Epischura lacustris, was found in two and seventeen bottles in
Experiment 1 and 2, respectively, its presence did not vary among treatments (one-way
ANOVA: both ps > 0.05) and the prey density in initial samples and control treatments
did not differ significantly (Student's t-test: all ps > 0.05) in both experiments. Hence we

combined data from the initial prey densities to those from the control treatments, and

73

used the average as an estimate of initial prey density in each experiment to increase
statistical power. We also removed one 25 umol In2 S" replicate and one 3 umol In2 S"
replicate in Experiment 1 and one 22 umol m'2 s" replicate in Experiment 2 from
analyses, because an individual Bythotrephes was not found in these bottles at the end of
the experiments.

We tested the hypothesis that light intensity led to different final densities of
zooplankton using a one-way ANOVA and separated differences between specific
treatments using Tukey HSD tests (Zar 1999).We followed this by calculating estimated
clearance rates (F, L hour" predator") and ingestion rates (I, prey hour " predator ") and
(a) testing for variation in clearance and ingestion rates for each experiment by one-way

ANOVA and then (b) combining data from the two experiments to estimate a relationship

between ingestion rate and light intensity. Clearance rate was calculated as:

w

(4.1) F: W

 

where Nfis final prey density (individuals L") in a Bythotrephes bottle, 1V1. is average

initial prey density (individuals L"), V is the volume of the bottle (L), t is the duration of
the experiment (hours), and P is the number of Bythotrephes in the bottle (Bamstedt et al.

2000). Ingestion rate was calculated as:

(4.2) 1 = F191.

We used standard error propagation techniques (Meyer 1975) to calculate standard errors
associated with mean clearance and ingestion rates of each Bythotrephes treatment.

Propagated standard errors are reported with treatment means and used in a subsequent

74

analysis, but not included in ANOVAs of clearance and ingestion rates because the error
associated with the initial prey density was common among Bythotrephes treatments.

We pooled estimates of ingestion rates from both experiments to better describe
Bythotrephes consumption over a broad range of light intensity (0.06 - 170 mol m'2 5").
Using weighted non-linear least squares regression, the relationship between light

intensity and Bythotrephes ingestion rate was fit to the logistic model:

(4.3) 1 = ———-—

where a, b, and c are fitted parameter representing the maximum ingestion rate, the light
intensity at which ingestion rate is half of its maximum, and the exponential rate of
increase, respectively, and L is light intensity (umol In2 5"). Data used for the regression
were treatment means weighted with their respective propagated standard error
(Gurevitch and Hedges 1999).We chose this logistic model because, for visual predators,
ingestion rate is often constant over low light intensities and then increases to an
asymptote as light intensity increases (Fraser and Metcalfe 1997).We chose to use
ingestion rate rather than clearance rate in this analysis, because the prey densities used in
both experiments were greater than that at which Bythotrephes predation is saturated at

high light intensities (K. L. Pangle, unpublished data).

RESULTS
In both experiments, Daphnia mendotae was the only of the seven prey types in
the prey assemblage whose density was significantly reduced (Experiment 1, F143 =

19.53,p < 0.0]; Experiment 2, 173,49 = 5.51, p < 0.01; Figure 4.1). In Experiment 1,

75

Prey density (number bottle")

20-

Experiment 1

 

 

[:1 Control (170 umol m'2 s", 0 Bythotrephes)
Low (3 pmol m'2 s", 3 Bythotrephes)
m Mid (25 umol m"2 s", l Bythotrephes)
High (170 umol m‘2 s", 1 Bythotrephes)

 

4o-

30-

20-

 

Experiment 2

   

 

'2 Control (22 limo] m'2 s", O Bythotrephes)
Low (0.06 umol In"2 S", 3 Bythotrephes)
m Mid (3 umol m'2 5", 3 Bythotrephes)
High (22 pmol m'2 s", l Bythotrephes)

*

 
   

..
— El
m

\.

“sew“

\

   

 

 

  

D.m. B.l. Nauplii I.Diap. A. Diap. I.Diac. A. Diac.

Prey Species

Figure 4.1. Densities of different prey types in Bythotrephes and control treatments in
Experiments 1 and 2. In each experiment, bottles were exposed to different light
intensities and received 0, 1, or 3 Bythotrephes depending on the treatment (see figure
legends). Bars are means with standard errors and reflect total, not per capita predator
effect. Data from initial prey density are included in mean densities of the control
treatment. Prey types included Daphnia mendotae (D.m.), Bosmina longirostris (B.l.),
copepod nauplii, immature Diaptomus (I. Diap.), adult Diaptomus (A. Diap.), immature
Diacyclops thomasi (I. Diac.), and adult D. thomasi (A. Diac.). Asterisks denote
significant differences.

76

Bythotrephes clearance and ingestion rates of D. mendotae did not differ significantly
between the 22 and 175 umol m"2 3" treatments; however, rates in these treatments were
significantly greater than those in the 3 umol m'2 5" treatment (Table 4.2), with values
approximately 4 times greater in the highest than the lowest light intensity treatment. In
Experiment 2, Bythotrephes clearance and ingestion rates of D. mendotae were
significantly greater in the 22 mo] m'2 5" treatment than those in the 0.06 and 3 umol m'
2 5" treatments (Table 4.2). An additional analysis using data from only the lowest light
intensity treatment showed that clearance and ingestion rates at this light intensity did not
differ significantly from zero (Student's t-test: t = 0.50, p = 0.63, df = 8; t3 = 0.13, p =
0.90, df = 8, respectively) and had upper bounds of 0.05 L h" and 0.31 Daphnia h" in
95% confidence intervals.

We pooled ingestion rate data found in Experiments 1 and 2 to examine the
functional relationship between Bythotrephes consumption of D. mendotae and light
intensity. The combined data fit a logistic model well (r2 = 0.98; Figure 4.2) with a half-

saturation point at 6.8 umol rn'2 s" and maximum ingestion rate at 2.1 Daphnia h".

DISCUSSION
Light strongly influenced the predation rate of the invasive planktivore,
Bythotrephes on the cladoceran Daphnia mendotae, a prey type shown to be highly
selected for by Bythotrephes (Vanderploeg et al. 1993; Lehman and Branstrator 1995;
Dumitru et al. 2001). Consumption of Daphnia was not detectable under low light
intensity (< 1 umol m'2 s"), but increased with greater light intensity, and eventually

leveled off at high light intensity (> 100 umol m'2 s"; Figure 4.2). The steepness of the

77

Table 4.2. Bythotrephes clearance rates (L hr") and ingestion rates (number hr") of
Daphnia mendotae observed in Experiments 1 and 2 at different intensities (description
of light intensities in Table 1).

 

Experiment 1

Experiment 2

Light intensity treatment

 

Variable Low Mid High
Clearance rate 0.16 d: 0.03 z 0.58 i 0.11 y 0.59 :t 0.10y
Ingestion rate 0.56 :t 0.07 z 2.06 :l: 0.40 y 2.08 i 0.34 y
Clearance rate 0.00 i 0.02 z 0.07 i 0.02 z 0.22 i 0.09y
Ingestion rate —0.02 i 0.17 z 0.49 :t 0.15 z 1.66 :l: 0.46 y

 

Note: Values are means i propagated standard errors and were calculated using equations
1 and 2. Mean row values with different lowercase letters were significantly different as
determined by ANOVA and Tukey’s HSD tests (p < 0.05).

78

 

 

 

3'0 - 0 Experiment 1
0 Experiment 2
2.5 a __
_f'\
.H .
=03 2.0 d
.c:
g c
g 1.5 -
5
8 J
3 1.0
a
.9.
Q 0.5 -
.53
0.0 a
'0.5 I l I I 1
0.1 l 10 100 1000

Mean light intensity (umol In2 S")

Figure 4.2. Effect of light intensity on Bythotrephes ingestion rate. Data of ingestion rates
of Daphnia mendotae in Experiments 1 and 2 were combined to evaluate Bythotrephes
predation over a broad range of light intensities. Points are means with propagated
standard error bars that were calculated using equations 1 and 2. In this analysis, the
relationship between light intensity and ingestion rate was described using a logistic
regression model (y = 2.12 / [1 + (x / 6.79)"‘57]), shown by the fitted line.

79

increase in predation rate at intermediate light intensities is difficult to ascertain given our
experimental design; however, the increase depicted in Figure 4.2 is similar to the
increase Muirhead and Sprules (2003) found for the reaction distance of Bythotrephes to
prey over a similar range of light intensities. Whereas we did not observe consumption of
other prey at any light intensity, our ability to detect predation for some of these prey
may have been limited due to their low density, particularly Bosmina longirostris, which
is another preferred prey of Bythotrephes (Vanderploeg et al. 1993).

Previous studies indicate that invertebrate planktivores primarily use tactile, rather
than visual cues, as the primary means to detect prey (Pastorok 1981; Kimboe and Visser
1999). Exceptions include studies on predation by mysids (Ramcharan and Sprules 1986)
and notonectids (Dieguez and Gilbert 2003); however, in both of these cases, these
invertebrate planktivores still relied strongly on tactile means to detect prey and, in
contrast to Bythotrephes, their feeding was not substantially reduced in total darkness.
Thus, the strong dependence of Bythotrephes on light intensity found here is unique
among known aquatic invertebrates and is comparable to planktivorous fish predation
that is often strongly affected by light intensity. Indeed, zooplankton prey have evolved
behavior known as diel vertical migration to avoid overlap with fish in areas and at times
of high light intensity (DeMeester et al. 1998). Laboratory experiments (Chapter 1) and
field observations (Chapter 5) have shown that zooplankton prey also migrate vertically
in the presence of Bythotrephes, suggesting that Bythotrephes may be driving selection
for zooplankton to perceive its risk and respond behaviorally.

A previous laboratory study by Muirhead and Sprules (2003) that evaluated prey

detection by Bythotrephes as a function of light also predicted that Bythotrephes

80

predation rate decreased with decreasing light intensity, but the predicted light effect was
less than what we found in our experiments. Their predictions were derived from a model
of the rate of prey encountered by a single Bythotrephes given the density and swimming
velocity of Bythotrephes and prey, and Bythotrephes reaction distance, which is a
function of light intensity (details of the model and its application provided in Appendix
A). The model was parameterized with data from laboratory observations of tethered
Bythotrephes' detection of prey. Given the densities in Experiment 2, the number of D.
mendotae encountered by Bythotrephes is predicted by the model to decline 29.1 % as
light intensity changes from 3 to 0.06 pmol m’2 5" (Appendix B); in contrast, the
observed mean decline in ingestion rate was actually 100 %.

The discrepancy between our results and the model predictions may result from
the ability of D. mendotae to detect Bythotrephes, which is not included in the model.
Zooplankton prey are known to avoid predators through hydromechanical signals prior to
an attack (Heath 1993; Viitasalo et al. 1998), such that the ability of prey to detect
predators is not dependent on light. To test the role of predator detection by Daphnia on
Bythotrephes predation rate at low light, we can apply models developed by Tiselius and
Jonsson (1990) parameterized with data from our study and other studies of Daphnia and
Bythotrephes (Brewer et al. 1999; Muirhead and Sprules 2003). These models predict the
hydro-mechanical signals produced by Bythotrephes and the distance from Bythotrephes
that D. mendotae would be able to detect these signals (details of the model and its
application provided in Appendix C). Results of the hydro-mechanical models show that
the predicted distance Bythotrephes is able to perceive prey is less than the predicted

distance that D. mendotae can detect Bythotrephes at low light intensity, but greater at the

81

mid light intensity. The model predictions suggest that D. mendotae may be able to
perceive Bythotrephes before being attacked at light intensities less than 2.7 umol m'2 s",
which could account for the dramatic decrease in predation observed, and the discrepancy
with the predation rate prediction by the Muirhead and Sprules (2003) model. These
idealized models provide a qualitative explanation for the patterns observed in our study,
however more research is needed on prey detection ability and how it influences
predation by Bythotrephes on D. mendotae. Building on previous studies that have
emphasized the importance of prey escape capabilities (Brenner et al. 1978; Heath 1993;
Viitasalo et al. 1998), our results suggest that light-mediated changes to the relative
ability of prey and predator to detect each other can substantially affect predation rate.

Light-dependent predation has implications for predicting Bythotrephes' impact
on zooplankton communities. Our results suggest that Bythotrephes effect could be
strongly influenced by the depth and time at which Bythotrephes and zooplankton
overlap. For example, there are periods in Lake Michigan in which the only overlap in
the vertical distribution of D. mendotae and Bythotrephes occurs at night (Chapter 5);
thus, Bythotrephes ability to consume D. mendotae would be greatly reduced.
Bythotrephes is also known to migrate downward itself when facing fish predation
(Straile and Halbich 2000). In lakes with high fish predation, predation by Bythotrephes
that are forced to inhabit deeper, darker regions of the water column would also be
greatly reduced (Lehman and Caceres 1993; Palmer et al. 2001).

The spread of Bythotrephes is of concern to management of lake ecosystems
given their potential impacts. Understanding factors controlling the establishment of

Bythotrephes in a lake can aid in identifying lakes to focus management efforts

82

(Muirhead and MacIsaac 2005). Our results suggest that the success of Bythotrephes, in
terms of its predation, would increase in lakes with increased water transparency; indeed,
this relationship has been observed in nature. MacIsaac et al. (2000) found that high
secchi depth values, a surrogate for high water transparency, was a strong predictor of
Bythotrephes presence in lakes within Bythotrephes native range. In addition, Branstrator
et al. (2006) reported a similar finding in Minnesota lakes that were recently invaded by
Bythotrephes. These studies propose different explanations for why Bythotrephes were
found more often in clear lakes. MacIsaac et al. (2000) suggest that high transparency
was indicative of systems with low productivity that could not support high densities of
planktivorous fish thereby providing Bythotrephes a release from both a predator and a
competitor. Branstrator et al. (2006) suggest that high transparency was not the primary
reason for the establishment of Bythotrephes, but rather an indicator that the systems also
have a warm, low-light, mid-depth water layer. This layer may be beneficial to
Bythotrephes populations by acting as a visual refuge from warrnwater fish and a thermal
refuge from coldwater fish (Yan et al. 2001). Based on our results, we propose that light-
dependent predation by Bythotrephes should be added to the list of potential factors
contributing to its establishment in clear lakes.

There is evidence that the type of predation observed in this study may be general
to other predatory cladocera. An experiment using Polyphemus pediculus found its
predation varied with light intensity and was not detectable in the absence of light (A. T.
Packard, unpublished data). A large, complex eye like the one Bythotrephes and
Polyphemus use to detect prey is a trait common to Cercopagis pengoi, Leptodora kindtii,

and ponids (Rivier 1998). The maintenance of a large eye is costly because it makes

83

predatory cladocera more visible to fish (Branstrator and H01] 2000). This large cost must
be associated with a large benefit, suggesting that the eye plays an important role in
detecting prey. In addition, the distribution of predatory cladocera in natural systems
indicates light influences their prey consumption, as they primarily inhabit well-lit areas,
such as the littoral zone and the epi- and metalimnion of the pelagic (Palmer et al. 2001;
Benoit et al. 2002). Future studies of light-dependent predation in other predatory
cladocera may aid in better understanding the role these organisms play in aquatic

communities and their relative success in different systems.

84

CHAPTER 5

LARGE NON-CONSUMPTIVE EFFECTS OF AN INVASIVE PREDATOR ON
ZOOPLANKTON POPULATION GROWTH RATE

ABSTRACT
We evaluated the non-consumptive effects of an invasive invertebrate predator,
Bythotrephes longimanus, on zooplankton prey populations in Lakes Michigan and Eric.
Field data taken at multiple dates and locations in both systems indicated that prey
species, Daphnia mendotae, Daphnia retrocurva, and Bosmina longirostris, inhabited
deeper portions of the water column as Bythotrephes biomass increased, possibly as an
avoidance response to predation. This induced migration reduces predation risk but also
can reduce birth rate due to exposure to cooler temperatures. We estimated the non-
consumptive (i.e., resulting from reduced birth rate) and lethal (i.e., consumptive) effects
of Bythotrephes on D. mendotae and B. longirostris. These estimates used diel field
survey data of the vertical gradient of zooplankton prey density, Bythotrephes density,
light intensity, and temperature with growth and predation rate models derived from
laboratory studies. Results indicate that non-consumptive effects played a substantial role
in the net effect of Bythotrephes on several prey population growth rates in the field, with
non-consumptive effects on the same order of magnitude as or greater (up to 10 fold)
than lethal effects. Our results further indicate that invasive species can have strong non-
consumptive, behaviorally-based effects, despite short evolutionary coexistence with prey

species.

85

INTRODUCTION

Predators can have large effects on prey distribution and dynamics. However,
what has recently become clearer is that predation itself is only one of a suite of effects
that a predator has on its prey. Predators induce changes in prey phenotype, including
behavioral, morphological, and life-history traits, in taxa as disparate as bacteria and
ungulates (reviewed in Lima 1998; Tollrian and Harvell 1999). Whereas such phenotypic
responses of prey can reduce predation risk, they are typically associated with a cost of
reduced growth rate (Werner et al. 1983; Peckarsky et al. 1993; Diehl and Eklov 1995;
reviewed in Peacor and Werner 2004), which can affect prey and predator abundance and
dynamics (e.g., Abrams 1995; Luttbeg and Schmitz 2000; Bolker et al. 2003). Further,
these induced effects on prey phenotype can ripple through a community in the form of
trait-mediated indirect interactions (Turner and Mittelbach 1990; Werner and Peacor
2003; Schmitz et al. 2004, also termed interaction modifications, Wootton 1994).
Therefore, “non-consumptive” predator effects may play a large role in ecological
systems, and improved understanding of their role may be critical to building predictive
ecological theory.

Empirical studies on non-consumptive effects of predators on prey have focused
almost exclusively on responses to individual prey grth rate and fecundity (Peacor and
Werner 2004). Recently, researchers have used mesocosms to demonstrate non-
consumptive effects of a predator on population-level responses in a damsel bug-pea
aphid system (Nelson et al. 2004) and in a Chaoborus-Daphnia system (Boeing et al.
2005). These analyses provide evidence that non-consumptive effects contribute to the

net effect of a predator on prey populations. Nevertheless, little is known of either

86

population-level non-consumptive effects or the relative importance of lethal and non-
consumptive effects in natural (uncontrolled) field environments.

We quantified both the lethal and non-consumptive effects of an invasive
invertebrate predator, Bythotrephes longimanus Leydig, on native zooplankton
populations in the Laurentian Great Lakes. Prior studies suggest that Bythotrephes has
impacted the pelagic ecosystem of the Laurentian Great Lakes (Lehman and Caceres
1993; Barbiero and Tuchman 2004) and smaller surrounding lakes (Hoffman et al. 2001;
Boudreau and Yan 2003). For example, the arrival of Bythotrephes coincided with loss
and reduction in density of several zooplankton species in Lakes Michigan (Lehman and
Caceres 1993) and Eric (0. E. Johannsson and D. M. Graham, unpublished data). Further,
laboratory experiments show that some zooplankton prey modify their behavior in the
presence of Bythotrephes by migrating to deeper, colder regions of experimental water
columns (Chapter 1), whereas a field study indicates that Daphnia mendotae Birge
vertical distribution deepened in Lake Michigan after the invasion of Bythotrephes
(Lehman and Caceres 1993). These patterns are indicative of an anti-predation strategy
commonly observed in zooplankton(G1iwicz 1986; Dodson 1988; reviewed in
DeMeester et al. 1998). If indeed zooplankton prey typically respond to Bythotrephes in
the field by inhabiting colder environments, such induced behavior may substantially
reduce prey population growth rate (a non-consumptive effect).

We focused our study on zooplankton species in Lakes Michigan and Eric that are
common, preferred prey items for Bythotrephes (Vanderploeg et al. 1993; Schulz and

Yurista 1999) in each lake. Our goals were to establish if Bythotrephes induce changes in

87

prey vertical distribution, and then to estimate the consequences of these changes on

population growth rate relative to lethal (consumptive) effects.

METHODS
We performed two field surveys denoted the “extensive field survey” and the
“intensive field survey”. In the former, we sampled the vertical distribution of
zooplankton during the day over a broad range of locations and dates to examine the
potential influence of Bythotrephes on prey vertical position. The intensive field survey
sampled fewer locations, but at more frequent depth and time intervals. The data from
these latter surveys were used with growth and predation models to estimate the

magnitude of Bythotrephes’ lethal and non-consumptive effects.

Influence of Bythotrephes on zooplankton daytime vertical distribution

Extensive surveys of Lakes Michigan and Eric were performed to evaluate the
influence of Bythotrephes on the percentage of prey inhabiting either the surface
epilimnion or the deeper, colder hypolimnion. Lake Michigan prey species included
Daphnia mendotae and Bosmina longirostris Muller, whereas Lake Erie prey species
were Daphnia retrocurva Forbes and B. longirostris. We hypothesized that prey should
be at lower risk in the hypolimnion, because Bythotrephes is a visually-orienting predator
that mainly inhabits the epi- and metalimnion (Muirhead and Sprules 2003, K. L. Pangle
et al., unpublished data). Therefore, we predicted that the percentage of prey occupying
the epilimnion would decrease, and the percentage occupying the hypolimnion would

increase as Bythotrephes biomass increased. We also examined whether abiotic

88

environmental factors influenced prey vertical distribution and therefore predation by
Bythotrephes. Specifically we evaluated the effects of epilimnion depth and secchi depth
(water clarity), because both affect the light intensity of the epilimnion and predation of
zooplankton.

Sampling was conducted during five different years and in multiple lake basins
(Table 5.1). Prior to each sampling event, thermal stratification of the water column was
determined using a submersible sensor. For Lake Michigan, sampling events consisted of
collecting 1 m3 of lake water from the center of the epilimnion, the center of the
metalimnion, and from the hypolimnion 10 m below the bottom of the metalimnion.
Water was collected using a diaphragm pump system and then filtered through a 64-um-
mesh zooplankton net on the deck of the research vessel. Samples collected with the
diaphragm pump system were similar to those taken by typical net sampling (K. L.
Pangle, unpublished data). Lake Erie samples were collected using a 110-um mesh
closing plankton net, 3 m long with a 0.5 m diameter opening. The net was towed through
each stratum, cinched at the desired depth using a secondary line, and then retrieved.
Secchi (20-cm black and white disc) depth was measured during each Lake Erie sampling
event. Samples were preserved in a buffered sugar-fonnalin solution and were subdivided
prior to enumeration with a Henson-Stempler pipette after gentle but thorough mixing.
Adult and juvenile zooplankton were identified to species and genus, respectively. For
each Lake Michigan sample, at least 600 individuals were counted, whereas for each
Lake Erie sample, at least 400 individuals were counted, with at least 100 individuals of
the major groups included, or if animals were scarce, 20 % of the sample was counted. For

all samples, Bythotrephes body lengths were measured using a drawing tube and digitizer

89

(Roff and Hopcroft 1986), and Bythotrephes biomass was calculated using a length-
weight relationship (Yan and Pawson 1998).

Percentage of prey in the epi- and hypolimnion (%), and Bythotrephes biomass
(mg/m2) were calculated using densities from each stratum weighted by the stratum
thickness. Statistical analysis of the effect of Bythotrephes biomass on these percentage
variables was performed for each zooplankton prey species using ordinary least squares
regression. A further analysis was performed that included epilimnion depth, secchi
depth, and Bythotrephes biomass as independent variables in a multiple linear regression
model to evaluate their effects on the percentage of B. longirostris and D. retrocurva
occupying the Lake Erie epilimnion. Prior to analysis, percentages were arcsine
transformed, and Bythotrephes biomass was log transformed to normalize distributions of

dependent and independent variables (Zar 1999).

Estimates of Lethal and Non-consumptive Effects of Bythotrephes

We estimated the non-consumptive and lethal effects of Bythotrephes on prey
population growth rate using data from the intensive field survey and the following
model of per capita prey population growth rate:

(5.1) —1—-d—1!=b—Ab-d—m
N dt

where b is the per capita birth rate in the absence of a phenotypic response, A is the
proportional reduction in per capita birth rate due to the phenotypic response to predation
risk (sensu Peacor and Werner 2004), dis the per capita consumption of prey by
Bythotrephes, and m is background per capita death rate due to other sources. This

representation allowed us to express the non-consumptive (Ab) and lethal (d) effect of

90

Bythotrephes with a common currency. Data requirements limited our analysis to species
that were most abundant, which included Daphnia mendotae from one survey in Lake
Michigan and Bosmina longirostris from two surveys in Lake Erie. We next describe the
intensive field surveys used to collect the necessary stage-specific data, and the
mathematical model, used to estimate Ab.

Non-consumptive effects of Bythotrephes were estimated using data from the
Lake Michigan intensive survey conducted on 3 and 4 August 2004 at a 60-m deep site
located approximately 10 km west of Muskegon, Michigan (Table 5.1). Four sampling
profiles were carried out over a 24-hr period, and profiles were initiated consecutively at
1400, 2100, 0100, and 0900 h. A sampling profile consisted of collecting 1 m3 of lake
water from five different depths ranging from 4 m to 40 m corresponding with the centers
of the epilimnion, metalimnion, hypolimnion, and the transitions between them. Water
was collected using the same methods described in the Lake Michigan extensive field
survey. Non-consumptive effects in Lake Erie were estimated from samplings collected
on 30 July and 8 September 1997 in Long Point Bay (Table 4.1). On each date,
Bythotrephes and zooplankton prey were collected at midday and midnight from 10
different depths that were distributed through the epi-, meta-, and hypolimnion. A
propeller-style pump was used to collect 250 L of lake water from each depth, and then
water was filtered through a 64-micron mesh net. Samples were processed and counted in
the same fashion as described for the extensive survey.

Field-derived estimates of D. mendotae and B. longirostris fecundity were
combined with known developmental rate-temperature relationship to estimate per capita

birth rate (b) using the egg-ratio method (Palheimo 1974) according to the equation:

91

0000:6800 00:30: 05 9 08—0 000002 0:03 3033 mo :0 £33002 0:80.30 m 00 08.5000 them EBA mg; 3 3:00 wfiafiwm...

 

 

m 33 Jazz» - 0:3 mm 3:095 cm» 2.69% a? 2.8303 0cm 033
m 32 ,:0£:0E0m - 33 E 3svmbm omw 2:003 om: 93:08 0cm 00.3
N wag .:0n::0a0m - Hmsws< mm 3:5ch L w 239$ omv A3553 0tm 0013
N woo“ Jmaws< - 33. mm 3.66am cow Z..oo.v~ om: £8808 0cm 0x04
3 33 :0nE0E0m - 0:3 voém 3..No.w~ can 2.593 cm: 490m “Eon w:o.d 0tm 0013
w M33 5:800 - 0:3.

2 $2 5:800 - 0:3. mm 3:3.8 0cm Z..mm..\.m cm:

0 go— .:0£:0E0m - 0:3. 683 0cm 00:3
n voom £3800 - 0:3. o: Birmm cow ZEN: am: 3:83 Emma—0:2 0073
m voom emsws< oo 3% Em cow Z..m r: omv 3:53 53:32 00:3
0 VOON .HmSWDdN l OCDH WV 3:N©.WN 00w 2:0NL w om? ASHDOmv GNWEO—wxé Dv—NQ

3:00 wfiasam S0» .0588 wfiafiam AEV 5Q0Q 033mg: 093:5 A3008 839$

 

.9030 20m 0cm 00?)— 0:0 :mwfiocz 00:3 8.: 00:0: 0:: 0:03:02 wfiacam ._.m 2an

92

ln(E/N +1)
D(T)

(5.2) b =
E is the density of eggs (eggs/m3), N is the density of prey (individuals/m3), and D(T) is
the egg developmental duration (d), which is a function of water temperature (T, °C). The
densities of eggs and individuals were averaged across sampling profiles and thermal
strata, weighted by the thickness of each thermal layer, to derive estimates of E and N,
respectively. D(T) has been derived for D. mendotae (Edmonson and Litt 1982) and B.
longirostris (Hanasato and Yasuno 1985) and can be estimated using the equations:

(5.3) D(T) =1/(O.00041-T2+ 0.0108 - T—0.0163)

and

(5.4) D(T) = 3.102 - 0.261 -1n(T2 ),

respectively. We integrated across the differing temperatures experienced by zooplankton
to derive a single estimate of D(T) for each survey (Hoffman et al. 2001) using the
equation:

(5.5) D(T) = Z D(T)pwp
p

where D(T)p is the mean developmental rate for each profile (p), and wp is a weighting

factor equal to the fraction of a 24 hour day each profile represented. For Lake Michigan,

the daytime profiles (0900 and 1400 h) and the nighttime profile (0100 h) were assigned
values of wp = 0.30 and wp = 0.20, respectively. The 2100 h sampling profile captured the

day-night transition during which D. mendotae migrated. Based on field observations of
D. mendotae vertical migration (K. L. Pangle et al., unpublished data) and Daphnia

swimming speed (Dawidowicz and Loose 1992), this sampling profile was assigned a

93

value of W], = 0.33. For Lake Erie, daytime and nighttime profiles were assigned values of
wp = 0.60 and wp = 0.40, respectively, for the first survey, and W], = 0.53 and wp = 0.47,

respectively, for the second survey, based on sunrise and sunset. D(T)p was in turn

calculated by averaging over the temperature experienced at different depths for each

profile

(5-6) D(T)p = ZD(T)Z,pfz,p
Z

where j}, p is the proportion of eggs relative to the total number of eggs in the water

column for depth 2 and profile p, and D(T);p is the estimated egg developmental duration

using Equation 3 or 4 given the temperature at depth 2 for profile p.

The non-consumptive effect of the predator (Ab in Equation 1) was calculated as
the difference between the estimated per capita population birth rate using the vertical
distributions of prey when Bythotrephes was present and when Bythotrephes was absent.
The latter distributions varied considerably for both species, thus providing us a range of
birth rates in the absence of Bythotrephes. We used the deepest, the average, and the
shallowest observed prey distributions in Bythotrephes’ absence from our extensive
surveys to determine a range of non-consumptive effects.

Lethal effects associated with Bythotrephes predation, d, were estimated using
known light-dependent consumption rates combined with the observed predator-prey
spatial overlap from the intensive field survey. We used predation rates derived from a
series of laboratory experiments in which Bythotrephes preyed on Daphnia at natural
prey densities in 4-L containers (slowly rotating in a 1500-L incubator) over a large range

of light intensities (K. L. Pangle et al., unpublished data). Note that these predation rates

94

are consistent with those in a previous study that used a bioenergetics approach (Y urista
and Schulz 1995). Vanderploeg et al. (1993) found that Bythotrephes consumed species
of Daphnia and Bosmina at a similar rate, so we used the same predation rates to estimate
B. longirostris consumed by Bythotrephes. However, Bythotrephes are likely better able
to detect D. mendotae than B. longirostris due to Daphnia's larger size (Confer et al.
1978), making our predation estimates for B. longirostris conservatively high. The per
capita rate of predation by a single Bythotrephes (a, prey consumed - m3 ° Bythotrephes'l '
d!) was estimated as:

8.196E — 03
1 + (L/24.96)‘ ”’34

 

(5.7) a =

where L is light intensity (umol - m'2 - s'l). The per capita (of prey) predation rate at a

given sampling profile (d , /d) was calculated as a weighted average:

aNPw
zzzz

 

(5.8) d =2
P NW
22

2

where az was predation rate at the depth 2 given its midpoint light intensity, Nz and P2

were the prey and Bythotrephes density (number/m3), respectively, at depth 2, and wz (m)

was the height of the water column represented by depth 2. For the Lake Michigan

sampling profile representing the day-night transition (2100 H),we used the highest light
intensity observed during the transition to calculate dp, thus making our estimate
conservatively high. The daily predation rate, considered the overall lethal effect of
Bythotrephes, was calculated as the arithmetic mean of the alps using the same weights for

sampling profiles as Equation 5.

95

RESULTS
Influence of Bythotrephes 0n daytime zooplankton vertical distribution

In Lake Michigan, Bythotrephes biomass had a significant negative influence on
percentages of both Daphnia mendotae and Bosmina longirostris in the epilimnion (B.
longirostris: Fm = 8.92, p = 0.01, ,2 = 0.41; D. mendotae: Fm = 9.65, p = 0.01, r2 =
0.45) and had a significant positive influence on percentages in the hypolimnion (B.
longirostris: Fl,l3 = 13.97, p < 0.01, r2 = 0.52; D. mendotae: Fm = 16.97, p < 0.01, r2 =
0.59). When Bythotrephes was absent from the pelagic community in Lake Michigan,
both prey species primarily inhabited the upper strata of the water column, with a greater
percentage of both prey species found in the epilimnion than in the hypolimnion (Figure
5.1). With high Bythotrephes biomass (i.e., > 200 mg/mz), prey vertical distribution
shifted downward and as much as 97 % of D. mendotae and 74 % of B. longirostris
inhabited the hypolimnion. This was particularly dramatic for D. mendotae, which were
almost entirely absent from the epilimnion at high Bythotrephes biomass.

In Lake Erie (Figure 5.2), Bythotrephes also had a significant negative influence
on percentages of prey, in this case B. longirostris and Daphnia retrocurva, in the
epilimnion (B. longirostris: 17152 = 7.84, p = 0.01, r2 = 0.13; D. retrocurva: F 1,40 = 4.95, p
= 0.03, r2 = 0.11) and had a significant positive influence on percentages in the
hypolimnion (B. longirostris: F152 = 5.47, p = 0.02, r2 = 0.10; D. retrocurva: F 1,12 = 8.23,
p < 0.01, r2 = 0.17). Indeed, both prey species in Lake Erie were almost entirely found in
the hypolimnion at high Bythotrephes biomass (i.e., > 300 mg/mz), and a decline in the
percentage of prey in the epilimnion started at 10 mg Bythotrephes/m2. There was greater

variation among sampling events in Lake Erie than in Lake Michigan, particularly when

96

.90: 003 5350 0330:: 05 3 8353va 05 300.050: 90 93% .8393 3000 :33? 00:000.: 0330: 07.30::
33>» .80: 0033333: :o 00:2: 0:0 80: 0088.35: :o 0038.0: 323000: 80:: 0:0 00:5 .ANEEEV 03:83 0033:0933
mo “Enema 0 :0>o 5033032 00304 :0 83:09»: 0:: -E0 05 3 A: .3 03330:: .Q 0:: A0 83 3.53:3:2 .m we 0m0::00:0m ._.m 03ME

ANEEEV 000803 .3§E.%=3 0.03%:0303

 

8:. 8m 8m 2: o 8:. 8m com 2: o
_ _ l . w , . N2.” 0: i . . .o
owoumk .53: 8.2L: E -8

5.338.2qu -

 

(%) 90111111110qu
9111 at afieruaorod

 

 

(%) uonrurqldo
9111 at ofieiuooad

3.0 n m: . :5 u N: . 8
5.0 n a .3: u S E .30. So u m .80 u 2 E E. 2:

 

 

 

 

00:30:05 StimuQ 0.3330023 :EEMSN

97

.0x8 003 b33038“ .9323

0330:: 0:83 0:0 $5 05 E 23395me 0:000:30 3v £288.20 98 3b mafim .8395. £000 GEES @3830 8:52 mmmbfla
3:3 4300 008082052: no @0303 can 3% 00803305 no wogofia 32300on 000:: 08 8:5 .ANEREV $0805 023000353
mo E0€Sw 0 00>o 0cm 831:0 cog—0%: was -E0 08 E G RC 9238:? .Q 0:0 6 .8 umhuehwzs .m «o 0w0§080m Nam 053m

ANEREV 9.0805 watcfimwzfi uuxmvhofkm

con coo cow oov com com o3 o con coo com oov 8m com 03 o

p -

 

 

 

 
 

 

....a- ... fio
N—HO” k 0 O OH.O“ 0..
a N n 0’ . . N‘s . O ”0‘ 1 ON U: 10
Sovm mmwuoim . No.onm,$.mu~m E . .. M m
. . ,8 mm
m m
r 1 CW 1.0%
0
u m.
. -a mq
- . . o .u - o2
. dd
. ".lu w
. mm
T m. 9
O. \N m.
:6an . ... . ‘2an Mom m/xw.
modumwaeuov: ... . um. . ”NS ..2:
3v :5 x: k E

 

 

 

92303000 EE‘QWQ “.385:ng 055.3%

98

Bythotrephes biomass was low (Figure 5.2); when Bythotrephes was absent from the
pelagic community, the percentage of B. longirostris and D. retrocurva ranged from 0 to
100 % in both the epi- and hypolimnion.

Abiotic environmental factors also influenced prey vertical distribution. In
particular, a significantly (p < 0.05) greater percentage of B. longirostris occupied the
epilimnion in Lake Erie as both epilimnion depth increased and secchi depth deceased. In
the multiple linear regression model, the magnitude of the Bythotrephes effect on the
percentage of B. longirostris occupying the epilimnion increased, a predicted trend;
however, this change was not significant (p = 0.18). There was no effect of epilimnion
depth and secchi depth on the percentage of D. retrocurva occupying the epilimnion (p >
0.05).

The vertical distribution of Bythotrephes remained constant throughout the Lake
Michigan intensive field survey, while a strong migratory behavior was observed for D.
mendotae (Figure 5.3). During the middle of the day, D. mendotae were distributed deep
in the water column, at dusk they had begun to move into upper vertical strata and
warmer temperatures, until at night, D. mendotae were found only in the upper stratum.
At dawn (0900 h), D. mendotae migrated back downward as light levels increased. This
migration resulted in a change in temperature to which prey were exposed (colder during
the day, Figure 5.3) and the degree of vertical overlap with Bythotrephes, which was high
during the nighttime but low during the day (Figure 5.3).

The vertical distribution of B. longirostris in the Lake Erie intensive field surveys
was below that of Bythotrephes during the day and nighttime sampling profiles (Figure

5.4). In the first survey, B. longirostris were found almost entirely in the hypolimnion at

99

.820 Bflwo—oaofio
E Eton“: 8m .8an mam—Q83 58 mo 50mm $023. Bow “Ema—BE gwfiomz 3:3 05 9.96 @ 9:28 333 Bu («0 582980
Em: 98 28809:»: >362: can .35 3.7.me wmfivhofiam 98 3:: unomv 353sz .Q mo noun—amaze Homto> .m.m oSwE

GOV 2338958 83>» Eb 8388 some E £26365 mo owficoeom

vm on 2 N— w v ow ow ow on o cw o0 ov om o ow cc ow cm o ow ow ov ON 0

_ — _ _ r p p _ . p p h . p . p . h p _ _

Em: ....... mqumhoikm ED... Ar
939388. I . SEESQ IOI.

 

(w) qldeo

 

 

 

 
 
 

 

 

 

 

 

A3 A3 Q as
o8 80 8e com o :88 :85 :88 f5:

 

A_-m.N-E._oEE 36:85 Emi mEEEmm mo 08:.

100

. . . -2 -1
Time ofsampling Light mtensrty (umol-m -s )

 

 

 

 

1200 h 2300 h 0 100 200 300 400 500
(a) + Bosmina (b) (c) — Temperature
WV“ Bythotrephes ....... Light .
5 < 4
E 10 r
5
8'
D 15 .
20 i
e =
25 . - . , , , . .

 

 

 

 

O 20 40 60 80 0 20 40 60 80 8 12 16 20 24
Percentage of individuals in each stratum (%) Water termperature (°C)

. . . -2 -1
L ht t l- -
Time of sampling 1g m ensrty (umo m s )

 

 

1300 h 0000 h 9 190 290 390 4%) 500
0 d .
( ) (e) + Bosmina (0 Temperature

 

' ...Vu Bythotrephes ....... Liglni......°

1

Depth (m)

 

 

 

 

 

 

25. . . . . . . . .. ...,
O 20 40 60 8O 0 20 40 60 801012141618 20

Percentage of individuals in each stratum (%) Water termperature (°C)

 

Figure 5.4. Vertical distribution of B. longirostris (solid line) and Bythotrephes (dashed
line), and mid-day temperature and light conditions of the water column during the first
and second Lake Erie intensive field surveys (a-c and d-e, respectively). Sampling
profiles were conducted near midday (a,d) and midnight (b,e) during both surveys.

101

midday and midnight, while Bythotrephes were found in the metalimnion at the same
times. In the second survey, B. longirostris were found in the hypolimnion, and
Bythotrephes were found in the epilimnion at midday and midnight. In both Lake Erie
surveys, vertical overlap between predator and prey was low and changed little between
sampling profiles. Overall, Bythotrephes biomass during intensive surveys was in the

midrange of Bythotrephes densities seen during extensive surveys (Figures 5.1 and 5.2).

Estimates of Lethal and Non-consumptive Effects of Bythotrephes

In Lake Michigan, the estimated lethal effect (d) of Bythotrephes on D. mendotae
population growth rate was 0.013 /d. Estimates of the non-consumptive effect (Ab) of
Bythotrephes on D. mendotae was 0, 0.038, and 0.174 /d, respectively, for scenarios that
used the deepest, average, and shallowest observed prey distributions in the absence of
Bythotrephes (Figure 5.5). In Lake Erie, d was estimated as 0.006 and 0.007 /d in the first
and second surveys, respectively (Figure 5.5). Estimates of non-consumptive effects in
the first survey were 0, 0.015, and 0.061 /d, respectively, based on three depth scenarios,
and in the second survey were 0, 0.024, and 0.079 /d, respectively. These estimates were

similar to those produced independently using a somatic grth rate model (Appendix

D).

DISCUSSION
This study demonstrates that non-consumptive effects can contribute strongly to,
and even predominate, the net effects of predators on prey population growth rate in a

natural setting. Variability in estimated effects among our different scenarios shows how

102

 

0.20 ‘ (a) —O— Lethal effect
---v"- Nonlethalefi‘ect v

0.15 i

0.10 ‘

Lake Michigan

predator effect (/d)

0.05 *

 

 

0.00
0.15 i

0.12 ‘

0.09 ~

0.06 ~

Lake Erie - First survey
predator effect (/d)

0.03 r

0.00 ‘4 . I l l r 2

0.15 ‘

 

 

0.12 ~
0.09 . ..v

0.06 4

Lake Erie - Second survey
predator effect (/d)

l ..... -. ..

0.03 <

 

 

 

a

V l I T I l

0.00 1
0 20 40 60 80 100

 

Percentage of prey in the epilimnion
Figure 5.5. Estimated lethal and non-consumptive effects of Bythotrephes on D.
mendotae in Lake Michigan (a) and B. longirostris in Lake Erie (b, c). Non-consumptive
effects (dotted line with open triangles) were estimated based on three different scenarios
(see Methods section) that assumed different prey distribution in Bythotrephes’ absence
to reflect the natural variation. This range is represented on the x-axis as the percentage
of the prey species in the epilimnion in Bythotrephes’ absence for the different scenarios.

103

the importance of non-consumptive effects may fluctuate over time. For example, at
times at which prey occupy deep waters independent of the presence of Bythotrephes,
Bythotrephes will not affect prey position, and non-consumptive effects are necessarily
absent. At other times, prey vertical migration appeared to be driven entirely by
Bythotrephes ’ presence and, in this scenario, non-consumptive effects were estimated to
be up to 10 times greater than lethal effects. Whereas the non-consumptive effect of
Bythotrephes may range from low to very large values in time and space, the relatively
frequent distribution of prey high in the water column in Bythotrephes absence indicates
that larger values are common.

If planktivorous fish biomass positively correlates with Bythotrephes biomass
then fish may underlie or contribute to the observed prey vertical distribution. We
therefore examined the correlation between planktivorous fish biomass and Bythotrephes
biomass, and found that they do not positively correlate. In particular, in Lake Michigan
during July and August 2004, acoustic estimates of fish biomass were relatively constant
in offshore areas where both Bythotrephes biomass and prey vertical distribution
fluctuated greatly (D. M. Krueger, unpublished data). Low variation in Lake Michigan
offshore fish biomass was also observed between May and September 1987 (Brandt et al.
1991) and is considered a general phenomenon (S. A. Pothoven and C. P. Madenjian,
personal communications). Other lines of evidence further support that Bythotrephes,
rather than fish, caused the vertical migration of zooplankton prey. Following the
introduction of Bythotrephes in Lake Michigan, zooplankton vertical distribution
deepened (Lehman and Caceres 1993), but the biomass of alewife, the primary species of

planktivorous fish, remained relatively constant (Madenjian et al. 2002). Further, the

104

underlying mechanism is supported by laboratory experiments, in which zooplankton
prey have been shown to respond strongly to Bythotrephes kairomones by migrating
downward in experimental columns (Chapter 1).

Because Bythotrephes consume prey at higher rates near the surface, we consider
if predation could underlie the change in prey distribution observed in Bythotrephes'
presence, rather than a behavioral mechanism. We believe this alternative is not plausible
because: (1) Bythotrephes has been demonstrated experimentally to induce strong vertical
migration in zooplankton (including D. mendotae and B. longirostris) in a manner
identical to that reported here (Chapters 1 and 2); (2) prey vertical migration was
observed in our intensive field surveys, particularly in Lake Michigan, where prey
inhabited very different regions of the water column during the day, night, and
transitional periods (Figure 5.3); (3) analysis of the extensive survey data (Appendix E)
indicated that absolute prey abundance in the hypolimnion increased with Bythotrephes
biomass, which is more indicative of a shift in prey habitat use than predation.

We have argued that the non-consumptive effect of Bythotrephes on prey birth rate
was a result of inhabiting lower temperatures. It is possible that different resource levels
experienced due to migration could also affect growth rate. However, results from
bioassay experiments showed that the grth of D. mendotae did not vary significantly
when cultured in Lake Michigan water taken from each of the thermal layers and held at
the same temperature (Chapter 3). Further, other studies have found that vertical
migration affects zooplankton birth rate primarily through differences in temperatures
and not from food resource differences (Winder et al. 2003; Park et al. 2004). It is also

possible that natural variation in resource levels could affect prey grth rate over time

105

and space and therefore confound comparisons of growth rate in the presence and
absence of Bythotrephes. This factor does not affect our results, because we compared the
predicted prey population growth rate in the presence and absence of Bythotrephes at the
same location and time.

Our results extend previous studies that indicate a strong contribution of non-
consumptive effects to the net interaction between predator and prey. Previous work has
primarily evaluated predator-induced reductions in prey somatic grth as a surrogate for
population grth (e. g. Diehl and Eklov 1995; reviewed in Peacor and Werner 2004).
However, although effects on somatic grth rate may translate into change in
population dynamics, it is also necessary to directly evaluate non-consumptive predator
effects at the population level (Nelson et al. 2004; Boeing et al. 2005) if we were to make
robust comparisons with lethal predator effects. Further, thus far, studies at the
population level have been performed in controlled mesocosm settings. It is essential to
translate experimental findings to patterns in natural settings to improve our
understanding of the hierarchy of processes that regulate ecosystems (Carpenter and
Kitchell 1988; Levin 1992; Schmitz 2005). By working in a setting as large and complex
as the Great Lakes, our approach is novel and necessarily very different than that used in
laboratory and mesocosm studies (e.g., rather than measure densities in a factorially
designed experiment, we made inferences from established relationships of growth and
predation). Overall, our results provide a demonstration that non-consumptive effects can
strongly influence population-level responses, even exerting greater effect on prey

populations than do lethal effects in large-scale, in natural settings.

106

The results of our study have implications to how biologists assess the impact of
Bythotrephes on zooplankton prey populations and the Great Lakes pelagic community
and the impact of predator-prey interactions in general. Non-consumptive effects can
introduce nonlinearities into predator-prey functional relationships, which could result in
a small number of predators having a disproportionately larger effect on prey density
(Abrams 1995; Peacor and Werner 2004), and could affect the stability of predator-prey
interactions (Luttbeg and Schmitz 2000). Furthermore, non-consumptive effects may
initiate trait-mediated indirect interactions (TMIIs) within food webs because changes in
the fitness and population size of responding prey may in turn affect interactions of the
responding prey with other species (reviewed in Werner and Peacor 2003; Schmitz et al.
2004). For example, other herbivorous zooplankton species less preferred by
Bythotrephes (V anderploeg et al. 1993 Schulz and Yurista 1999) may respond differently
to Bythotrephes predation risk, thus altering competitive interactions (Dawidowicz and
Wielanier 2004). In addition, changes in the assemblage and densities of zooplankton
may affect other predators, like young-of-year fish, which can clearly have cascading
effects on the food web.

Finally, the evolutionary history of species interactions may strongly influence the
phenotypic responses of a prey species to a predator (Abrams 2000; Trussell 2000;
Cousyn et al. 2001) and therefore non-consumptive effects. Whereas there are few studies
of phenotypic responses of prey to invasive predators, the magnitude of the response has
been shown to vary greatly from no change at all (Pearl et al. 2003) to strong changes
(McIntosh and Townsend 1995). Further, the magnitude of the response has been shown

to change through time (Cousyn et al. 2001). Given the variation in the phenotypic

107

response, the magnitude of the non-consumptive effect will vary accordingly. In the case
studied here, native zooplankton prey in the Great Lakes respond strongly to the presence
of Bythotrephes after a relatively short evolutionary history (approximately 20 years). Is
this predator-induced response a result of rapid evolution, or did prey have this ability
prior to the invasion of Bythotrephes due to adaptation to native predators? Exploration
of this evolutionary question could shed light into changes in non-consumptive effects on
Great Lakes zooplankton over the last two decades and further our understanding of the

non—consumptive effects of invasive species in general.

108

CHAPTER 6

INTERACTIONS BETWEEN LAKE MICHIGAN COMPETITORS MEDIATED BY
THE BEHAVIOR OF A SHARED PREY

ABSTRACT
We modeled the interactions between Bythotrephes longimanus, Daphnia mendotae, and
age-0 alewife (Alosa pseudoharengus) in Lake Michigan to evaluate how consideration
of trait-mediated indirect interactions between Bythotrephes and alewife would change
predictions of the impact Bythotrephes has on alewife. We used two models that were
linked together and ran simultaneously. The first model estimated the effects of
Bythotrephes on the population dynamics and vertical distribution of Daphnia over the
course of a season (June 1 - October 31). The second model estimated consumption and
growth of age-0 alewife over the same time period. Daily Daphnia density from the first
model was added to a pool of prey items in the second model that were available for
alewife to encounter and consume. Model predictions depended strongly on whether or
not the NCEs of Bythotrephes were included For example, when only CEs of
Bythotrephes were considered, predicted mean length of alewife entering winter was 110
mm and predicted over-winter survival was 87 %. When NCEs were added, both by
including a cost of migration on Daphnia population growth rate and considering the
reduction in overlap between fish and Daphnia resulting from migration, predicted mean
length of alewife entering winter was 72 mm and predicted over-winter survival was 19
%. Thus, model results suggest that Bythotrephes primarily affects fish grth through

trait-mediated indirect effects caused by induced changes in prey behavior. These

109

findings provide an example in a natural system of how interactions between competitors

can be strongly influenced through adaptive behavioral responses of shared prey.

INTRODUCTION

The detrimental effects of human disturbance on ecosystems (Jackson et al. 2001;
Essington et al. 2006) combined with the growing demand on ecosystem services
(Bennett and Balvanera 2007) have exacerbated the need for sound management of our
natural resources. One response has been a paradigm shift within managing from a single
species context to managing within a multi-species or ecosystem-based context
(Browman et al. 2004; Pikitch et al. 2004). The motivation here is to better account for
the complex effects humans may have on ecosystems by considering ecosystem-level
processes such as indirect interactions, feedback loops, and shifts between alternative
regimes (Curry 2004).

Ecosystem-based management relies heavily on food web models (Christensen
and Walters 2004). Food web models link species populations in the ecosystem based on
predator-prey interactions. These interactions are typically represented by consumptions
rates; however, there is an increasing awareness that predators can affect prey through
means other than direct consumption in ways that may profoundly influence food web
processes (Abrams 1995; Werner and Peacor 2003). For example, prey may respond
adaptively to predators by modifying their traits (reviewed in Harvell 1990; Sih 1997;
Lima 1998; Tollrian and Harvell 1999). Trait modifications are important to predator-
prey interactions because they generally reduce the vulnerability of prey and because they

come at a cost to prey grth rate (Werner et al 1983; Peckarsky et al. 1993; Diehl and

110

Eklov 1995; Nakaoka 2000; reviewed in Peacor and Werner 2004). Theory indicates that
non-consumptive effects of predators may be strong enough to affect prey and predator
abundance and dynamics (Ives and Dobson 1987; Abrams 1992, 1995; Bolker et al.
2003). Further, trait modifications may affect how predators and prey interact with other
species in an ecosystem, a phenomenon referred to as trait-mediated indirect interactions
(Turner and Mittlebach 1990; Werner and Peacor 2003; Schmitz et al. 2004; Preisser et
a1. 2005; also termed interaction modifications, Wootton 1994).

Recently, some aspects of predator-induced trait modifications have begun to be
incorporated into food web models. Most frequently incorporated are prey vulnerability
functions, in which prey become less vulnerable to predators as predation mortality
increases (Dill et al. 2003; Christensen and Walters 2004; Garrott et al. 2007).
Consideration for prey vulnerability has been shown to improve model predictions of
food web dynamics in large, complex ecosystems (Harvey et al. 2003; Araujo et al.
2006). However, in most applied food web models, decreased vulnerability does not
come with an associated cost to prey. Further, the vulnerability of a prey to one predator
is generally considered independent of the prey's vulnerability to another predator
(though see Walters and Christiansen 2007). It is not clear how exclusion of these
interactions may affect food web model predictions, but based on ecological theory and
experimental findings, the effects may be profound.

Here we present a model of consumptive and non-consumptive interactions
occurring in Lake Michigan pelagic ecosystem. We focus specifically on the competitors,
age-0 alewife (AIosa pseudoharengus) and Bythotrephes longimanus, and their shared

prey resource, zooplankton. Alewives are a dominant forage fish in Lake Michigan and

111

the primary food resource for salmonids in Lake Michigan (Madenjian et al. 2002).
Bythotrephes is a predatory cladoceran that was introduced into Lake Michigan via
ballast water in the 19805. Bythotrephes has been credited with changing the zooplankton
community (Lehman and Caceres 1993), and is a particular concern to biologists, because
its effect on zooplankton may affect the survival of age-0 alewife (Shuter and Mason
2003), whose recruitment into the adult population depends in part on the availability of
zooplankton as a food resource (Hook et al. 2007).

In addition to direct consumption of zooplankton, experiments and field
observations indicate that Bythotrephes induces changes in zooplankton behavior. As
Bythotrephes density increases, zooplankton migrate to deep-water refugia during the
daytime to avoid being eaten. Bythotrephes-induced migration may indirectly affect
alewife in two ways. First, zooplankton migration causes reductions in zooplankton
population grth rate due to inhabiting cooler water temperatures, ultimately leaving
less zooplankton for alewife to consume. Second, zooplankton migration reduces
zooplankton spatial overlap with alewives, which inhabit shallow regions of the water
column during their first year of life.

Here we constructed a food web model that includes the contribution of induced
zooplankton migration to the net predicted impact of Bythotrephes on age-0 alewife. Our
goal was to elucidate the significance of non-consumptive effects to food web predictions
by comparing their importance relative to the traditionally considered consumptive
effects. This exercise also aids in better understanding the impact of Bythotrephes in

invaded ecosystems.

112

METHODS

We used two models that were linked together and run simultaneously over a
simulated growing season (Figure 6.1). The first model predicted seasonal zooplankton
population dynamics as a function of abiotic conditions and interactions between
zooplankton and Bythotrephes. We focused specifically on one zooplankton species,
Daphnia mendotae, which is the most dominant species during the summer (Barbiero and
Tuchman 2001) and thought to be an important food resource to Bythotrephes (Lehman
and Caceres) and age-0 alewife (Stewart and Binkowski 1986). The second model
estimated the prey consumption and growth of age—0 alewife over the same time period.
Daily Daphnia density from the first model was added daily to a pool of other prey items
in the second model that were available for alewife to encounter and consume. We ran
the models under environmental conditions observed in Lake Michigan at a site 20 km
offshore of Muskegon, Michigan (Figure 6.2).We chose this location because interactions
among Bythotrephes, alewife, and Daphnia have been observed there, and it is where the
off-shore-abundant Bythotrephes and the near-shore-abundant alewife have the strongest
spatial overlap. In addition, densities of age-0 alewife at this site do not achieve high
enough levels to substantially affect the density or behavior of Daphnia (K. L. Pangle,
unpublished data), thus simplifying the interactions required in the model.

We ran the models under eight different scenarios of indirect effects of
Bythotrephes on alewife that occur from all possible combinations of the presence and
absence of (1) direct consumption of Daphnia, (2) reduction in Daphnia birth rate due to
induced vertical migration, (3) reduction in vertical overlap between Daphnia and alewife

due to induced vertical migration (Table 6.1). We compared outcomes of the scenarios

113

.550 some 8 woo—a: 03 >05 Bo: 98 33m £5 5 wow: £0on 95 65 mo £380:an 4.0 oSwE

6608 £3on 95 E58 858:?
cosmeacoo hum ergo—4c. mamas—smog SSEGQ
coumfismcoo
55 t/
53 floobo 03035
All 33%qu can
oEm=m>< amwxmmhofiam

6.3on k

£362

114

 

24 _ (a) —— Epilimnion
--------- Hypolimnion

   

20-

16~

12-

..............
...........
.........
.........
.........
,-
,-
.r
.u
.n
.-
u
.-
..-
.n
.-
.I'
-
..-
..0'
,-
,-
.I‘

Water temperature (°C)

 

16 - (b)

14-

12‘

Day length (hours)

10-

 

16 - (c)

12-

B ythotrephes density
(number per m'3)

 

 

 

0 I I I I I I I

180 200 220 240 260 280 300

Julian date

Figure 6.2. Seasonal environmental variables simulated in the model, including water
temperature (panel a), day length (panel b), and Bythotrephes density (panel c).

115

 

OO><><><O><
O><O><O><><
XOOOXXX
vamxotxoo

O O 0 fl

 

amigo .892, £503-95:ch E =28:qu 88 5.5 ESEQQ E gauged EEEQQ mo 538328 885 2.8:on

 

.0936? so “.2358er mo mwoobo 68:3: Eobb€ ADV @2535 use 00 wows—2: 35 85:08 1.608 Emma 4.0 £an

116

based on predicted size of age-0 alewife over the simulated growing season and predicted

over-winter survival of age-0 alewife given the size they attained at the start of winter.

Daphnia population dynamics model

The dynamics of a Daphnia population (N) were modeled using predicted daily
per capita birth rate (1)) and death rate (d) (i.e., no net immigration or emigration) and the
equation:

(6.1) fl = (b —d)N
dt

Birth and death rates depended on seasonal changes in water temperature, day length, and
Bythotrephes density (Figure 6.2). These parameters were derived from observations in
Lake Michigan, and Bythotrephes density was not dynamically linked to Daphnia
density.

Daphnia birth and death rates vary over the course of the day due to Daphnia diel
vertical migration, changes in sunlight intensity (Chapter 5), and the density of Daphnia.
To account for this variation, we estimated birth and death rates for the daytime,

crepuscular time, and nighttime, and then averaged the birth and death rates across the
different periods, weighted by the portion of time (w,) represented by each period. In the

model, the daytime was the fraction of the day between sunrise and sunset (Figure 6.2),
the crepuscular time was the two hours prior to sunrise and the two hours following
sunset, and the nighttime was the remaining hours of a 24-hr day.

Daphnia in the model inhabited either the hypolimnion or the epilimnion and the
fraction of the Daphnia population in each stratum was dependent on Bythotrephes

density (B) and time of day (Chapter 5). During the daytime period, the fraction of

117

Daphnia in the hypolimnion (pdayjrypo ) and in the epilimnion (pdayflpi or 1- pday’hypo) was
estimated using the equation:

(6.2) pday'hypo = sin(0.04651 + 0.1485010ge(3.2 B+1))
which was derived from field observations of the change in the fraction of Daphnia in the
hypolimnion over a gradient of Bythotrephes densities (Chapter 5). During the
crepuscular period, the relative portion of Daphnia in the hypolimnion (pday’hypo ) was 70
% of the density during the day period, while the other 30 % was added to the density of
Daphnia in the epilimnion. This change in distribution is based on observations of
Daphnia migration in Lake Michigan (Chapter 5). During the night period, all Daphnia
were in the epilimnion.

Daphnia per capita birth rate (b, day") was derived by averaging across the

different birth rates experienced in a day due to changing temperature using the equation:
(6.3) b = gbtwt

where b, is the birth rate predicted for time period t and w, is the time-weighting factor.

For each time period, b, was predicted using the equation:

 

pth
(6.4) 1952122,th 1— K
2

where p2 was the portion of Daphnia in strata z , bz was the maximum birth rate of

Daphnia in strata z, and K was carrying capacity (K; 11,000 individuals m'3). Maximum

birth rate and carrying capacity was derived from both experimental and observational

118

data (K. L. Pangle, unpublished data). Birth rate of strata 2 was calculated as function of

water temperature (T, °C) using the equation:
(6.5) b, = 0.0002 T, 2 + 0.0051 T, - 0.0077

where T2 is the mid-point temperature of strata z. The equation was derived

experimentally (Edmonson and Litt 1982) and assumed that Daphnia fecundity was
independent of temperature.
Deaths of Daphnia were due solely to predation by Bythotrephes, and Daphnia

were only vulnerable to Bythotrephes predation while in the epilimnion. The per capita
rate of predation by a single Bythotrephes (at, individuals consumed m'3 Bythotrephes'1

day") during time period t was modeled as:

0.008196

L — 1.034
+ I
24.96

where L, is the light intensity (umol rn'2 sec") typically observed at the mid-point depth

 

(6.6) a t =

 

of the epilimnion during time period t (Chapter 5). Equation 6 was derived
experimentally (Chapter 4). Light intensities used for the daytime, crepuscular time, and
nighttime were 600, 60 and 1 umol rn'2 sec", respectively. Per capita death rate of

Daphnia for each time period was calculated using the equation:

(6.7) dt = sz’tatB
Z

119

where p2,, was the portion of Daphnia in epilimnion during time period t and B is

Bythotrephes density. Per capita rate of predation for each time period was averaged,
weighted by period length, as in equation 3.

The density of Daphnia from July 1 to October 31 was solved numerically using
the Runge-Kutta (4,5) approximation. The initial density of Daphnia (1 individual m'3)
was estimated from field observations.
Alewife prey consumption and growth model
The daily consumption and growth of alewife was modeled from its hatch date to the start
of winter. The model was adapted from one presented in Fulford et al. (2006), which
modeled the same processes for age-0 yellow perch in Lake Michigan, and consisted of a
consumption sub-model that determined the number and type of prey eaten and a
bioenergetics sub-model that incorporated consumption into an energy budget to
determine somatic growth. In this model, alewife consumed seven prey types (Table 6.2)
at rates that depended on encounter rates, handling time, and alewife selectivity. Prey
types, except for Daphnia and Bythotrephes, was kept at constant, conservatively high
densities that were observed in Lake Michigan during periods in which Daphnia was
absent (K. L. Pangle, unpublished data). The lengths and weights of prey types were

averages observed in Lake Michigan (Fulford et al. 2006).

Consumption sub-model
Encounter rate (E), prey sec") of each prey type was a product of the swimming

speed and reaction area of alewife and the density of the prey type. Swimming speed

120

ER“ .35: E 3388 v.8 E58 05 E 828% than...

 

EE cmAHw How 3.3:an we 08%

 

 

 

 

we Robe “A

as omvd. é o com S m e e m
E. 80366 T ~86 :
omogx mg C Wm ad 2 ESEQQ
3%
8.0
. Ge 8:866 T |.1 +_ _
:5 ox v o 2 no 8% 655.3
....o

as 2A.: 5 to . . 5% a
as :vfi E #32 - its Ea mm _ a 0 8w 6. _ 0

2A.: “8 Nd . . Ego 9n
Svfi § 3 - .2 who + be 9.3. mm o a. o O? 6. _ 0
So 2 .o No 89. £632
3.5-1:. @323 _.o m _ .o ooooc eased

3330922 333028 3% Emmy? 5 3:5 £w=2 Lon. “on “BEE bacon 35 >05

 

.EvoE 05 E com: 090 DE :23 mo mocmtflofiwso mo SEESm .N6 2an

121

(mm sec'l) was expressed as a function of wet weight (WW, g) and water temperature (T)

using the equation:

(6.8) SS [ WW

0.274
1,000,000]

ww, Temp = exp(0.056-T)

derived experimentally for alewife (Klumb et al. 2003). Reaction area (RA ,3 TL) was
expressed as a function of prey size and fish size using the equation:
( l2

(6.9) RA. = i 7: 05

“TL rad
2tan[ T%]
K /

 

 

 

where PL,- is the length (mm) of prey i, and rad“, is the angle (radians) of visual acuity

for a larvae of a given TL. This equation is commonly used for age-0 fish (Rose and

Cowan 1993, Letcher et al. 1996). We used the relationship between rad and TL:
(6.10) rad = 0.0167 exp(9.14 — 2.4ln(TL)+ 0.2291n(TL)2)

which was found experimentally for bluegill from 10 to 150 mm TL and swimming in
clear water (Breck and Gitter 1983) and has been applied to multiple species of fish
(Letcher et al. 1996).

The probability that alewife consume an encountered prey (QLTL, unitless) is

calculated as:

ai,iTLN

(611) Q
TL zzajJLN j

 

122

where a)“, is the selectivity of a fish of length k for prey type i and N, is the density of

prey type i (Lockwood 1998). The relationships between a and TL for each prey type are
not well known for alewife, so they were derived from experiments using age-0 yellow
perch (Fulford et a1. 2006, Table 6.2), except for Bythotrephes. Yellow perch have similar
predatory behavior relative to alewife during early life-history stages (Miller et al. 2007).
Alewife are incapable of consuming Bythotrephes until they reach approximately 50 mm
TL (Branstrator and Lehman 1996), and, after fish reach 50 mm TL, their selectivity on
Bythotrephes is similar to that on Daphnia (Pothoven et al. 2007). Consequently, the a
for Bythotrephes was kept at 0 for fish < 50 mm TL, and the same as that for Daphnia for

fish > 50 mm TL.

Handling time (see) was expressed as a function of PL (mm) and TL (mm) using

the equation:

7.0151-[111]
L

(6.12) H =exp 0.264-10

PL, TL
The equation was derived using yellow perch data from Mills et al. (1984) and Miller et
al. (1992). Miller et al. (1992) showed that the relationship between body size and
handling time of prey for yellow perch and alewife were very similar.

Consmnption rate (C, 11g day") was calculated as:
Z Ei, TLQi,TLP W i, TL

(6.13) C = ' ~DL
1+ZEi,TLQi,TLHPL,TL
l

 

123

where PW,- is the dry weight (11g) of prey i, HpLTL is the handling time (sec) for prey of
length (PL) for an alewife of a given TL, and DL is day length (sec). Consumption rate

(C, ug day") could not surpass a maximum level (C max) due to physiological constraints.

Cmax is a function of fish size and water temperature, and this relationship has been

specifically derived for alewife (Hanson 1997).

Bioenergetics sub-model

Consumption rate (C) was converted into daily growth using a bioenergetics
model developed specifically for alewife (Stewart and Binkowski. 1986; Hewett and
Stewart 1989). For positive growth to occur, energy gained through consumption needed
to surpass the total loss of mass due to respiration, egestion, excretion, and specific
dynamic action. Respiration, the amount of mass used by fish for routine metabolism was
a function of water temperature and body size (Post 1990). Egestion, the portion of
consumption that is not ingested (e. g., loss to sloppy feeding), is held constant at 16 % of
consumption. Excretion, the portion of ingested food that is not assimilated, is held
constant at 16 % of ingestion. Specific dynamic action, the portion of ingested food lost
to the metabolic demand of assimilation, is held constant at 17.5 % of ingestion. For the
calculations, we assumed that the energy density (Calories g'l) of alewife and its prey
were equal (Post 1990).

The mass gained (lost) due to consumption and the physiological processes was
added (subtracted) to fish's current weight to determine its weight for the next day. If the
change was positive, then the new weight was used to determine a new length. Total

length was converted to wet weight (WW, ug) using the equatiOns:

124

exp(- 6.7091 + 4.3157 . 010%)». 1,000,000 if TL < 43.4 mm
(6.14) WW =

exp(— 4.72 + 2.96 - 160%)» 1,000,000 if TL 2 43.4 mm
(Stewart and Binkowski 1986). Some functions in the model required the mass of fish in

terms of dry weight (DW, 11g). The dry-to-wet weight ratio is dependent on fish size, and

can be calculated using the equation:

(6.15) DW/WW = -0.00001-TL2 + 0.0026-TL + 0.0945
derived by fitting data published for alewife (Lantry and O'Gorman. 2007; Hewett and
Stewart 1989) using nonlinear least-squares regression.

The model began as alewife hatched over multiple dates. The distribution of hatch
dates was based on observations from Lake Michigan (Hook et a1. 2007), with mean date
of July 1 and the first and last hatches occurring on June 20 and July 31 days,
respectively. The initial total length (TL, mm) of alewife on each hatch date was 6 mm.
This value represents the size at which alewife typically begin to feed (Heinrich 1981),

and was used by Hook et al. (2007) to assign hatch dates in Lake Michigan.

Scenario simulations and comparisons

In order to run the model under the different scenarios, the indirect effects of
Bythotrephes on alewife were systematically removed (Table 6.1). Direct consumption of
Daphnia by Bythotrephes was removed by setting per capita rate of predation by a single
Bythotrephes (a) to zero. The reduction in Daphnia birth rate due to induced vertical

migration was removed by setting the water temperature of the hypolimnion equal to that

125

of the epilimnion. The reduction in vertical overlap between Daphnia and alewife due to
induced vertical migration was removed by allowing alewife to consume Daphnia in the
epilimnion and hypolimnion.

Four variables were evaluated to determine the influence of NCEs, including daily
Daphnia density and the size and growth rate of alewife over the course of the season,
and alewife over-winter survival. Alewife overwinter survival was estimated using the
distribution of final lengths of alewife at the start of winter and a size-dependent over-

winter survival rate observed in Lake Michigan (Hook et al. 2007).

Sensitivity analysis

We evaluated the sensitivity of the model to selected parameter values by varying
each parameter by i 50 % and measuring the effect of this change on predicted over-
winter survival of alewife (Christensen and Walters 2004). Sensitivity analysis was only
carried out under "full effect" scenario, in which Bythotrephes had the largest effect on

alewife.

RESULTS
Daphnia densities initially increased in all scenarios, with the greatest rates
observed for scenarios in which Bythotrephes had no effect on Daphnia growth rate, and
the lowest for the “full effect” scenario (Figure 6.3a). Daphnia densities in the “no effect”
scenario were near carrying capacity by day 250, while Daphnia densities in scenarios
that included direct consumption tended to peak around day 260 and decline thereafter, as

the water temperature began to decline and Bythotrephes density continued to increase.

126

Daphnia density available to Total Daphnia density

Daphnia density available to

alewife (individuals per m'3) (individuals per m'3)

alewife (individuals per m'3)

 

   
  

 

 

  
  
   

 

 

  

 

 

 

20000 - (a) With and without reduced overlap
No effect
16000 ‘ ------------------ Direct consumption
—————— Reduced birth rate
—- -------- Direct consumption /
_ /
12000 and reduced birth /
rate y .......
8000 « /
/
/
4000 4 /
/
20000 ‘ (b) No direct consumption
No NCEs
16000 T .................. Reduced birth rate ............
—————— Reduced overlap
12000 ~ — -------- Reduced birth rate
and overlap
8000 .
/’—_--‘,,:“=-‘-=
4000 - / _. /~/
0 .. __ /
20000 ‘ (c) Direct consumption
16000 — No NCES
.................. Reduced birth rate
12000 d ------ Reduced overlap
— -------- Reduced birth rate
and overlap
8000 -
4000 4
O '1 ..., ......._...-—— ''''''''
I60 200 240 280 320
Julian day

Figure 6.3. Seasonal total density of Daphnia (panel a) and density of Daphnia available
for alewife to consume (panels b, c) under the eight model scenarios. Total Daphnia
density is not affected by reduced overlap with alewife, thus the eight model scenarios
are represented by four lines.

127

The density of Daphnia available from consumption by alewife followed similar
trajectories as total Daphnia density (Figure 6.3b and c), with the only differences
occurring between scenarios that included the overlap effect.

Alewife growth rate over the simulated season varied considerably among
scenarios. The “no effect” scenario had the greatest growth rate (Figure 6.4) with a mean
final length of 116 mm. In this scenario, the distribution of alewife sizes tightened as the
season progressed, leading to the lowest variation in final alewife length among the
scenarios (Figure 6.5). The “full effect” scenario had the lowest growth rate (Figure 6.4)
with mean final length with a mean final length of 70 mm. Contrary to the "no effect"
scenario, the distribution of alewife sizes remained relatively constant in scenario 8,
leading to the highest variation in final alewife length among the scenarios (Figure 6.5).

Over-winter survival rate also varied considerably among the scenarios. As
expected, Scenario 1 had the highest over-winter survival at 93.1 %. The inclusion of
each indirect effect alone led to relatively small reductions in over-winter survival, with
direct consumption having the least effect (Figure 6.4). As effects were combined, over-
winter survival dropped steadily, with the lowest value, 18.7 %, obtained when all three
effects were included.

Sensitivity analysis showed that model predictions of over-winter survival in
scenario 8 were most greatly affected by changes in parameters associated with water
temperature, Daphnia birth rate, Bythotrephes density, and aleWife foraging (Table 6.3).
For parameters in the Daphnia population dynamics model, a 50% increase in the

epilirnnetic water temperature led to the greatest increase in over-winter survival. For

128

.motmcoom 3on Emma 05 581526 .88 88 3323 88 88 538w :38 dogma—Emma .vd 85wE

cage—8.6.28 885 corms—=88 88:. oz

 

5:868 88 58553638 88 5.3 oZ 5:038 88 5.5 «.2838 88 5.3 oz

 

 

 

 

Soto “onto Soto Soto Soto «cote Soto “onto
amigo 3:30 02 nah—:30 9830 02 3:30 @233 oZ 92.30 3:26 02

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3v

1dN

19.

1dc

1cm

1 do—
dd

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

E

1 Nd

1 vd

1 ed

1 ad

1 d.—

 

 

(%) [BAIMUS smurmalxo petogpatd

(Kep rad mm) 9113.! qworfi ueaw

129

 

     
  
  

    
  
  

 

 

 

   

 

 

 

 

‘20 J SCENARIO l: ‘ SCENARIO 2:
l 00 _ No effect Direct consumption
80 < I
60 '1 .4
40 4 '1
20 d a
0 T I I r I I T I I
120 -
SCENARIO 3: SCENARIO 4:
100 q Reduced birth rate - Reduced overlap
80 4 -
60 ~ -
E 40 ~
5 i
v
..z: 20 a I
E”
3 0 r f I I I I I I I
"g? '20 ‘ SCENARIO 5: ‘ SCENARIO 6:
3 100 - Reduced birth rate -« Direct consumption
3‘: and overlap and reduced overlap
E 80 4 T
<
60 — -
40 a -
20 - -
O I I I I I T I I I
120 - 7 I
SCENARIO 7: SCENARIO 8:
100 - Direct consumption u Full effect
and reduced birth rate
80 J -
6O - -
40 J ~
20 a a
0 I I I I I I I I I
160 200 240 280 320 160 200 240 280 320
Julian day

Figure 6.5. Lengths of alewife over the season that were hatched on the first (solid line),
mean (dotted line), and last (dashed line) hatch date under the eight model scenarios.

130

Table 6.3. Results of sensitivity analysis, in which the overwinter survival in the "full
effect" (original 18 %) was estimated after a d: 50 % in each parameter.

Overwinter survival (%)

 

Parameter with 50 % decrease with 50 % increase
Daphnia population dynamics model
epilimnion temperature (T ep,-) 16 86
hypolimnion temperature (Thypo) 16 43
Bythotrephes density (B) 59 16
Intercept of equation 2 19 19
Slope of equation 2 17 23
Daphnia carrying capacity (K) 18 19
epilirnnetic birth rate (bepi) 16 76
hypolimnetic birth rate (bhypo) 16 31
Bythotrephes daytime predation rate (aday) 42 16
Bythotrephes crepuscular predation rate(acrep) 25 17
Bythotrephes nighttime predation rate(a,,,-g;,,) 19 18

Alewife consumption and growth model

Initial Size of alewife (TL at hatch date) 0 46
Weight-length conversion (equation 8) 8 100
Dry weight - wet weight conversion 6 27
ODW/W W)

Rotifer density 0 35
Rotifer alpha value 23 16
Nauplii density 25 14
Nauplii selectivity coefficient 25 13
Cyclopoid density 21 17
Cyclopoid selectivity coefficient 22 16
Calanoid density 2 38
Calanoid selectivity coefficient 0 33
Bosmina density 3 31
Bosmina selectivity coefficient 18 19
Daphnia selectivity coefficient 17 20
Bythotrephes selectivity coefficient 15 22
Swimming speed (SS) 0 55
Angle Of visual acuity (rad) 71 0
Handling time (H) 51 9
Maximum consumption 19 19
Respiration rate 54 0
Egestion 38 2
Excretion 33 5
Specific dynamic action 42 O

 

131

parameters in the alewife consumption and grth model, a 50 % decrease in the angle of

visual acuity led to the greatest increase in over-winter survival.

DISCUSSION

This study demonstrates the significant role that non-consumptive interactions can play in
food webs. Namely, Bythotrephes was predicted to have a large effect on alewife only
when interactions in addition to direct consumption of a shared prey were considered. In
addition, the manner in which survival rate declined with added effects indicates that the
effects had complex, synergistic interactions that led to greater reductions in alewife
survival than would be predicted by sum of reductions of each effect alone. Such
interactions between CBS and NCES have been shown empirically (Peacor and Werner
2001) and result here due to the non-linearity of alewife growth and survival.

While this model is still a coarse abstraction of the processes occurring in Lake
Michigan, evidence suggests that this model produces patterns that are comparable to
those observed in Lake Michigan. For example, Pothoven et al., (2001) report several
years in which Daphnia population in offshore Lake Michigan have a gradual increase in
density at the start of the season, which Slows or reverses around day 250. In addition,
HOOk et al. (2007) found that in offshore Lake Michigan the length of age-O alewife
ranged from 35 to 100 mm, with a mean length of 66 mm. These field observations are
very Similar the distribution of lengths predicted when all three indirect effects of
Bythotrephes were considered (Figure 6.5).

Results of this study therefore indicate that consideration of NCES may better

capture the nature of food web interactions and thus better inform ecosystem-based

132

management that rely on food-web models. NCES between competitors, Bythotrephes
and alewife, have been largely excluded from food web model studies. Existing spatially-
explicit models may capture some of the interactions Shown to be important here (e. g.,
Walters et al. 1999); however, these models would only be appropriate for antipredator
behaviors involving Spatial movement and not for other behaviors, such as vigilance
(Lima 1998), or other traits modifications, such as changes in prey morphology and life-
history (Tollrian and Harvell 1999), that may also be important to indirect interactions
between competitors. One justification for the exclusion of NCES between competitors in
food web models follows an evolutionary argument, in that if two competitors have
foraging behaviors that induced a shared prey to modify the same trait, then intense
competition between the predators will ensue and natural selection would favor
difi‘erentiation of their foraging behavior (Walters and Christiansen 2007). Following this
argument, consideration for NCES between competitors in food webs that include
invasive species, such as Lake Michigan, may be particularly important given the lack of
co-evolution.

Explicitly considering NCES will likely be a challenging endeavor for ecosystem-
based management. Most studies of NCES have been conducted in controlled laboratory
and mesocosm settings, and their applications to large, complex ecosystems will
necessarily require novel approaches. The approach we took in this study was to first
develop a functional relationship that related predator density with trait induction by
prey. Second, we quantified the cost of trait induction at the prey population level. Third,

we considered the effect of trait inductions on interactions between the prey and other

133

Species in the ecosystem. The third requirement is a particularly severe departure from
conventional food web models that are intrinsically based on pair-wise interactions.

In the context of the Great Lakes ecosystem, the results of this study suggest that
Bythotrephes may be having detrimental effects on alewife recruitment. Indeed, the
arrival of Bythotrephes has coincided with the declines in alewife populations in Lake
Michigan and Lake Huron. It is diffith to ascertain whether these declines were caused
by Bythotrephes because other dramatic changes have also occurred during this time
period, such as the introduction of Dreissenid mussels and increases in salmon
populations. In addition, the model presented here is directly applicable only to the
offshore regions of Lake Michigan, and it is not well known how recruitment of alewife
in the offshore regions contributes to the total alewife population. Consideration of large-
scale Spatial distribution of age-0 alewife and Bythotrephes may thus provide further

insights into the impact on Bythotrephes on the Lake Michigan food.

134

APPENDIX A

A list of articles associated with the dissertation chapters. All chapters were
collaborative efforts with Scott Peacor. Chapter 5 also represents a collaborative effort

with Ora Johannsson.

Chapter one. Pangle, K. L., and S. D. Peacor. 2006. Non-lethal effect of the invasive

predator Bythotrephes longimanus on Daphnia mendotae. Freshwater Biology 51:
1070-1078.

Chapter two. Pangle, K. L., and S. D. Peacor. Lack of naivete' in Lake Michigan
zooplankton prey to risk posed by invasive predators. In prep.

Chapter three. Pangle, K. L., and S. D. Peacor. Temperature gradients, but no food
resource gradients, affect growth rate of Daphnia mendotae in Lake Michigan. In

prep.

Chapter four. Pangle, K. L., and S. D. Peacor. Light-dependent predation by the
invertebrate planktivore, Bythotrephes longimanus. Submitted to Canadian
Journal of Fisheries and Aquatic Sciences.

Chapter five. Pangle, K. L., S. D. Peacor, and O. E. Johannsson. 2007. Large non-
consumptive effects of an invasive invertebrate predator on zooplankton

population growth rate. Ecology 88: 402-412.

Chapter Six. Pangle, K. L., and S. D. Peacor. Interactions between Lake Michigan
competitors mediated by the behavior of a shared prey. In prep.

135

APPENDIX B

Gerritsen and Strickler (1977) developed a model that predicted how predators
encountered prey in aquatic systems, assuming that predator and prey: (1) are randomly
distributed; (2) swim in random directions; (3) swim at an average velocity; and (4)
encounter each other when prey are within a reaction Space around the predator. This
reaction Space is described by a solid angle (3) using the equation:

(3.1) s = sin6d6 d¢
where 6 is the polar angle and ¢ is the azimuthal angle. An encounter model was

adapted specifically for Bythotrephes-prey encounters by Muirhead and Sprules (2003).
In this model, the rate of prey encounter (Z, number h") by a single Bythotrephes is

estimated using the equation:

 

 

Zr.
2 2752

(3.2) Z=R2N I [sineJuzwz—zuvcoso d6 d¢
o o

where R is the reaction distance of the predator (m), N is prey density (number m3), and
u and v are the swimming velocities (m h'l) of the prey and predator, respectively. Using
a series of laboratory experiments, Muirhead and Sprules (2003) related Bythotrephes
reaction distance (mm) to light intensity (L, umol m'2 s']) according to the set of
equations:

(B3)

136

(6.61, L < 2

R =1 —7.69+36.97log10(L +1) — 21.8810g10(L +1)2 +3.93log10(L + I)3, 2 s L s 500
L9.995, L > 500

 

We estimated difference in the rate of D. mendotae encountered between the low and
mid-light treatments of Experiment 2 using this model given our experimental conditions.
From the literature, we found average swimming velocities (mm S") for second instar
Bythotrephes (16.03; Muirhead and Sprules 2003) and Daphnia mendotae (1.2; Gerritsen
1980). Model predictions were compared with difference between encounter rates
observed in Experiment 2 calculated using ingestion rates (I) and the equation:

(B4) 1 = N (1 — exp(—Z (T — 111))

where T (hours) is the total duration of the experiment and h (hours) is an experimentally-
derived estimate of the handling time associated with the consumption of individual
Daphnia by Bythotrephes (K. L. Pangle, unpublished data). This recursive equation
allowed us to estimate encounter rate for Bythotrephes exposed to a saturating level of

prey density (Hassel 1978).

137

APPENDIX C

A swimming planktivore produces water velocity gradients that can alert
zooplankton of its presence. Studies have shown that zooplankton are particularly
sensitive to the deformation rate (d, S") component of water velocity gradients caused by
a predator (Kierboe and Visser 1999; Kicrboe et al. 1999). Assuming that a planktivore
produces water velocity gradients that are similar to that of a sphere, with a radius, r
(em), and a velocity, v (cm s"), and assuming that zooplankton respond to a planktivore
when deformation rate surpasses some critical threshold (A, S"), the distance that

zooplankton can detect a planktivore (D, cm) can be calculated using the equation:

(C.1) D=r 3 1+ 1_§:_A_ forASa
4rA 3v 8r

 

Using this model, we estimated the distance that D. mendotae could detect Bythotrephes
using a critical threshold found for Daphnia (1.8 s", Brewer et al. 1999), Bythotrephes
swimming velocity (1.6 cm S"; Muirhead and Sprules 2003), and the mean thoracic

radius of Bythotrephes used in the experiment (0.15 cm).

138

APPENDIX D

We used a second method to estimate the non-consumptive effect (Ab) of the
predator on prey population growth rate. In this case, we used somatic growth rate as a
surrogate for population growth rate. For cladoceran species like D. mendotae and B.
longirostris, somatic growth rate correlates very strongly with population growth rate (r2
= 0.99, Lampert and Trubetskova 1996).We estimated somatic growth rate (SGR, /d) for
D. mendotae and B. longirostris using the equations:

(D.l) SGR = -0.0009 T2 + 0.0476 T- 0.1882

and

(D2) SGR = -0.0002 T 2 + 0.0020 T - 0.0506,

respectively, where T is water temperature (°C). These equations were derived by fitting
nonlinear models to published data on D. mendotae (Hall 1964) and B. longirostris

(Hanasato and Yasuno 1985). The mean somatic growth rate of each sampling profile
(SGRP) was calculated as:

ZSGR N w
2 Z 2
(13.3) SGRp = Z

 

NW
22
z

where SGRZ was the estimated somatic growth rate at the depth 2 given its midpoint
temperature, NZ was the prey density (number/m3) at depth 2, and wz (m) was the height

of the water column represented by depth z. SGRPS were averaged over the entire 24—h

period using the same weights as Equation 5 in the main text. The non-consumptive

effect of the predator was then calculated as the difference between the estimated somatic

139

growth rate using the vertical distributions of prey when Bythotrephes was present and
when Bythotrephes was absent. The former distributions came from the three intensive
surveys, while the latter distributions were the deepest, the average, and the shallowest
observed prey distributions in Bythotrephes absence from our extensive surveys.

This method yielded qualitatively similar estimates of the non-consumptive effect,
relative to the lethal effect, Of Bythotrephes on prey growth rate that of the more direct
method based on life history of organisms collected in the field. For D. mendotae in Lake
Michigan, the non-consumptive effect on somatic growth rate was estimated as 0, 0.101,
and 0.211 /d, respectively, based on our three scenarios in which prey were at the
deepest, the average, and the shallowest Observed prey distributions in Bythotrephes
absence from our extensive surveys. For B. longirostris in Lake Erie, estimates of non-
consmnptive effects in the first survey were 0, 0.048, and 0.105 /d, respectively, while

estimates were 0, 0.016, and 0.050 /d, respectively, in the second survey.

140

APPENDIX E

In the extensive field survey, Bythotrephes biomass had a negative influence on
the percentage of prey in the epilimnion and a positive influence on the percentage of
prey in the hypolimnion (Figs. 1 and 2).We interpreted these patterns as reflecting an
induced vertical migration that was an adaptive behavioral response to predation risk.
Here we evaluate an alternative possibility, namely that the observed changes in prey
distribution are due to predation by Bythotrephes; i.e. predation by Bythotrephes, which
was found primarily in the epilimnion, would also cause a reduction (increase) in the
percentage of prey in the epilimnion (hypolimnion).

Consider the abundance of prey Species in the hypolimnion. If Bythotrephes
predation, not induced prey migration, were responsible for changes in prey distribution,
then Bythotrephes would not affect the abundance of prey in the hypolimnion. Therefore,
there would not be a correlation between Bythotrephes biomass and prey density. In
contrast, if changes in prey distribution resulted from induced vertical migration, then we
would expect prey abundance in the hypolimnion to be positively correlated with
Bythotrephes biomass. Therefore a positive correlation would be consistent with our
interpretation of induced migration but not predation. Note that these predictions require
that there is no correlation between total prey abundance (in the water column) and
Bythotrephes biomass. Because we expect either consumption or induced vertical
migration to negatively affect prey abundance, the inference that a positive correlation
between Bythotrephes biomass and abundance in the hypolimnion resulted from induced

vertical migration is conservative. We used linear regression analysis to examine the

141

relationship between prey abundance in the hypolimnion and Bythotrephes biomass. Prior
to analysis, absolute prey abundance and Bythotrephes biomass were log-transformed to
normalize distributions of dependent and independent variables (Zar 1999). There was a
positive relationship between Bythotrephes biomass and the abundance of Daphnia
mendotae and Bosmina longirostris in the hypolimnion (D. mendotae: t1; = 2.02, p =
0.03; B. longirostris: (13 = 1.69, p = 0.05) in Lake Michigan, and of B. longirostris and D.
retrocurva abundance in the hypolimnion (B. longirostris: t5; = 2.37, p = 0.01; D.
retrocurva: I40 = 2.42, p = 0.01) in Lake Erie. Therefore, analysis of the absolute
abundance of Bythotrephes prey are consistent with our interpretation that the prey
relative abundance at different levels was affected by Bythotrephes-induced changes in

vertical migration rather than predation.

142

LITERATURE CITED

Abrams P. A. 1982. Functional responses of optimal foragers. American Naturalist 120:
382-390.

Abrams P. A. 1993. Why predation rate should not be proportional to predator density.
Ecology 74: 726-733.

Abrams, P. A. 1995. Implications of dynamically variable traits for identifying,
classifying, and measuring direct and indirect effects in ecological communities.
American Naturalist 146: 112-134.

Abrams, P. A. 2000. The evolution of predator-prey interactions: Theory and evidence.
Annual Review of Ecology and Systematics 31: 79-105.

Abrams, P. A. 2007. Defining and measuring the impact of dynamic traits on
interspecific interactions. Ecology 88: 2555-2562.

Aksnes, D. L. 2007. Evidence of visual constraints in large marine fish stocks.
Limnology and Oceanography 52: 198-203.

Ali, M. A. 1959. The ocular structure, retinonomotor, and photobehavioral response of
juvenile Pacific salmon. Canadian Journal of Biology 37: 965-996.

Bamstedt, U. D. J. Gifford, X. lrigoien, A. Atkinson, and M. Roman. 2000. Feeding. In
Zooplankton methodology manual. Edited by R. P. Harris, P. H. Wiebe, J. Lenz,
H. R. Skjoldal, and M. Huntley. Academic Press, New York, NY. pp. 297-400.

Barbiero, R. P., L. L. Schacht, M. A. DiMartino, and M. L. Tuchman. 2000. Effects of
vertical distribution of zooplankton on the estimation of abundance and
biovolume using deep and shallow tows. EPA Great Lakes Monitoring Report.

Barbiero, R. P., and M. L. Tuchman. 2001a. Results for the U.S. EPA’S biological open
water surveillance programs of the Laurentian Great Lakes: 11. Deep chlorophyll
maxima. Journal of Great Lakes Research 27: 155-166.

Barbiero, R. P., and M. L. Tuchman. 2001b. Results for the U.S. EPA's biological open
water surveillance programs of the Laurentian Great Lakes: III. Crustacean
zooplankton. Journal of Great Lakes Research 27: 167-184.

Barbiero R. P., and M. L. Tuchman. 2004. Changes in the crustacean communities of
Lakes Michigan, Huron, and Eric following the invasion of the predatory

cladoceran Bythotrephes longimanus. Canadian Journal of Fisheries and Aquatic
Sciences 61: 2111-2125.

143

Bennett, E. M., and P. Balvanera. 2007. The future of production systems in a globalized
world. Frontiers in Ecology and Evolution 5: 191-198.

Benoit, H. P., O. E. Johannsson, D. M. Warner, W. G. Sprules, and L. G. Rudstram.
2002. Assessing the impact of a recent predatory invader: the population
dynamics, vertical distribution, and potential prey of Cercopagis pengoi in Lake
Ontario. Limnology and Oceanography 47: 626-635.

Blackburn, T. M., P. Cassey, R. P. Duncan, K. L. Evans, and K. J. Gaston. 2004. Avian
extinction and mammalian introduction on oceanic islands. Science 305: 1995-
1958.

Boeing W. J ., B. Wissel, and C. W. Ramcharan. 2005. Costs and benefits of Daphnia
defense against Chaoborus in nature. Canadian Journal of Fisheries and Aquatic
Sciences 62: 1286-1294.

Bolker, B. M., M. Holyoak, V. Krivan, L. Rowe, and O. Schmitz. 2003. Connecting
theoretical and empirical studies of trait-mediated interactions. Ecology 84: 1101-
1114.

Bottrell, H. H. 1975. The relationship between temperature and duration of egg
development in some epiphytic Cladocera and Copepoda from the river Thames,
Reading, with a discussion of temperature functions. Oecologia 18: 63—84.

Boudreau, S. A., and N. D. Yan. 2003. The differing crustacean zooplankton
communities Of Canadian Shield lakes with and without the nonindigenous

zooplanktivore Bythotrephes longimanus. Canadian J oumal of Fisheries and
Aquatic Sciences 60: 1307-1313.

Brandt, S. B., D. M. Mason, E.V. Patrick, R. L. Argyle, L. Wells, P. A. Unger, and D. J.
Stewart. 1991. Acoustic measures of the abundance and size of pelagic
planktivores in Lake Michigan. Canadian Journal of Fisheries and Aquatic
Sciences 48: 894-908.

Branstrator, D. K., M. E. Brown, L. J. Shannon, M. Thabes, and K. Heimgartner. 2006.
Range expansion of Bythotrephes longimanus in North America: evaluating

habitat characteristics in the Spread of an exotic zooplankter. Biological Invasions
8: 1367-1379.

Branstrator, D. K., and C. M. H01]. 2000. Planktivory by bluegill (Lepomis macrochirus)
on Leptodora kindtii in a small North American lake. Hydrobiologia 437: 101-
106.

Branstrator, D. K., and J. T. Lehman. 1996. Evidence for predation by young-of-the-year
alewife and bloater chub on Bythotrephes cederstroemi in Lake Michigan. Journal
of Great Lakes Research 22: 917-924.

144

Breck, J. E., and M. J. Gitter. 1983. Effect of fish size on the reactive distance of bluegill

(Lepomis macrochirus) sunfish. Canadian Journal of Fisheries and Aquatic
Sciences 40: 162-167.

Brewer, M. C., P. Dawidowicz, and S. I. Dodson. 1999. Interactive effects of fish
kairomones and light on Daphnia escape behavior. Journal of Plankton Research
21: 1317-1335.

Brooks, J. L., and S. I. Dodson. 1965. Predation body size and composition of plankton.
Science 150: 288.

Browman, P., M. Cury, R. Hilbom, S. Jennings, H. K. Lotze, P. M. Mace, S. Murawski,
D. Pauly, M. Sissenwine, K. I. Stergiou, and D. Zeller. 2004. Perspectives on
ecosystem-based approaches to the management of marine resources. Marine
Ecology Progressive Series 274: 269—303.

Bruno, J. F ., J. D. F ridley, K. D. Bromberg, and M. D. Bertness. 2005. Insights into biotic
interactions from studies of species invasion. In D. F. Sax, J. J. Stachowicz, and S.
D. Gaines [eds], Species Invasions: Insights into Ecology, Evolution, and
Biogeography, Sinauer Associates, Sunderland, Massachusetts.

Bungartz B., and D. K. Branstrator. 2003. Morphological changes in Daphnia mendotae
in the chemical presence of Bythotrephes longimanus. Archiv fiir Hydrobiologie
158: 97-108.

Byers J. E., S. Reichard, J. M. Randall, I. M. Parker, C. S. Smith, W. M. Lonsdale, I. A.
E. Atkinson, T. R. Seastedt, M. Williamson, E. Chomesky, and D. Hayes. 2002.

Directing research to reduce the impacts of nonindigenous species. Conservation
Biology 16: 630-640.

Carpenter, S. R., and J. F. Kitchell. 1988. Consumer control of lake productivity.
Bioscience 38: 764-769.

Carpenter, S. R., and J. F. Kitchell. 1993. The trophic cascade in lakes. Cambridge
University Press.

Christensen, V., and C. J. Walters. 2004. Ecopath with Ecosim: methods, capabilities,
and limitations. Ecological Modeling 172: 109-139.

Clark, C. W., and D. A. Levy. 1988. Diel vertical migration by sockeye salmon and the
antipredation window. American Naturalist 131: 271-290.

Colton, T. F. 1987. Extending functional-response models to include a 2nd prey type - an
experiment test. Ecology 68: 900-912.

Confer, J. L., G. L. Howick, M. H. Corzette, S. L. Kramer, S. Fitzgibbon, and R.
Landesberg. 1978. Visual predation by planktivores. Oikos 31: 27-37.

145

Cousyn, C., L. De Meester, J. K. Colboume, L. Brendonck, D. Verschuren, and F.
Volckaert. 2001. Rapid, local adaptation of zooplankton behavior to changes in

predation pressure in the absence of neutral genetic changes. Proceedings of the
National Academy of Science (USA) 98: 6256-6260.

Creel, S., and D. Christianson. 2008. Relationships between direct predation and risk
effects. Trends in Ecology and Evolution 23: 194-201.

Dawidowicz, P., and C. J. Loose. 1992. Cost of swimming by Daphnia during diel
vertical migration. Limnology and Oceanography 37:665-669.

Dawidowicz, P., and M. Wielanier. 2004. Costs of predator avoidance reduce competitive
ability of Daphnia. Hydrobiologia 526: 165-169.

DeMeester, L., P. Dawidowicz, E. van G001 and C. J. Loose. 1998. Ecology and
evolution of predator-induced behavior of zooplankton: Depth selection behavior
and diel vertical migration, p. 160-177. In R. Tollrian and C. D. Harvell [eds],
The Ecology and Evolution of Induced Defenses. Princeton University Press,
New Jersey, USA.

DeMott, W. R., and B. J. Pape. 2005. Stoichiometry in an ecological context: testing for

links between Daphnia P-content, growth rate and habitat preference. Oecologia
142: 20-27.

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

Dieguez, M. C., and J. J. Gilbert. 2003. Predation by Buenoa macrotibialis (Insecta,
Hemiptera) on zooplankton: effect of light on selection and consumption on prey.
Journal of Plankton Research 25: 759-769.

Diehl, S., and P. Eklov. 1995. Effects of piscivore-mediated habitat use on resources,
diet, and growth of perch. Ecology 76: 1716-1726.

Dill, L. M., M. R. Heithaus, C. J. Walters. 2003 Behaviorally mediated indirect
interactions in marine communities and their conservation implications. Ecology
84:1151-1157.

Dodson, S. I. 1988. The ecological role of chemical stimuli for the zooplankton -
predator-avoidance behavior in Daphnia. Limnology and Oceanography 33:
1431-1439.

Dodson, S. I., and D. G. Frey. 2001. Cladocera and Branchipoda. pages 850-914 in eds. J.
H. Thorp and A. P. Covich. Ecology and Classification of North American
Freshwater Invertebrates. 2nd edition. Academic Press, San Diego, California.

146

Dorazio R. M., J. A. Bowers, and J. T. Lehman. 1987. Food-web manipulations influence
grazer control of phytoplankton growth rates in Lake Michigan. Journal of
Plankton Research 9: 891-899.

Drenner, R. W., J. R. Strickler, and W. J. O'Brien. 1978. Capture probability: the role of
zooplankter escape in the selective feeding of planktivorous fish. Journal of the
Fisheries Research Board of Canada 35: 1370-1373.

Dumitru, C., W. G. Sprules, and N. D. Yan. 2001. Impact of Bythotrephes longimanus on
zooplankton assemblages of Harp Lake, Canada: an assessment based on predator
consumption and prey production. Freshwater Biology 46: 241-251.

Edmonson, W. T., and A. H. Litt. 1982. Daphnia in Lake Washington. Limnology and
Oceanography 27: 272-293.

Ehrlich, K. F. 1974. Chemical changes during growth and starvation of herring larvae.
Pages 301-323 in J. H. S. Blaxter, editor. The early life history of fish. Springer-
Verlag, New York.

Elton, C. S. 1927. Animal ecology. Sidwick and Jackson, London.

Evans, M. S. 1988. Bythotrephes cederstroemi: its new appearance in Lake Michigan.
Journal of Great Lakes Research 14: 234-240.

F ahnenstiel, G. L., and D. Scavia. 1987. Dynamics of Lake Michigan phytoplankton: the
deep chlorophyll layer. Journal of Great Lakes Research 13: 285—295.

F ahnenstiel, G. L., R. A. Stone, M. J. McCormick, C. L. Schelske, and S. E. Lohrenz.
2000. Spring isothermal mixing in the Great Lakes: Evidence of nutrient
limitation and a nutrient-light interaction in a sub-optimal light environment.
Canadian Journal of Fisheries and Aquatic Sciences 57: 1901—1910.

Fraser, N. H. C., and N. B. Metcalfe. 1997. The costs of becoming nocturnal: feeding

efficiency in relation to light intensity in juvenile Atlantic salmon. Functional
Ecology 11: 385-391.

F ryxell, J. M., and P. Lundberg. 1997. Individual behavior and community dynamics.
Chapman and Hall, New York.

Fulford R. S., J. A. Rice, T. J. Miller, F. P. Binkowski, J. M. Dettmers, and B. Belonger.
2006. Foraging selectivity by larval yellow perch (Percaflavescens): implications
for understanding recruitment in small and large lakes. Canadian Journal of
Fisheries and Aquatic Sciences 63: 28-42.

Garcia, S. M., and K. L. Cochrance. 2003. Ecosystem approach to fisheries: a review of
implementation guidelines. ICES Journal of Marine Sciences 62: 311-318.

147

Garrott, R. A., J. E. Bruggeman, M. S. Becker, S. T. Kalinowski, and P. J. White. 2007.
Evaluating prey switching in wolf-ungulate systems. Ecological Applications 17:
1588-1597.

Gerristen, J. 1980. Adaptive response to encounter problems. In Evolution and ecology Of
zooplankton communities. Edited by W. C. Kerfoot. University Press New
England, Hanover, NH. pp. 52-62.

Gerristen, J ., and J. R. Strickler. 1997. Encounter probabilities and community structure
in zooplankton: a mathematical model. Journal of the Fisheries Research Board of
Canada 34: 73-82.

Gielbelhausen, B., and W. Lampert. 2001. Temperature reaction norms of Daphnia
magna: the effect of food concentration. Freshwater Biology 46: 281-289.

Gliwicz, M. Z. 1986. Predation and the evolution of vertical migration in zooplankton.
Nature 320: 746-748.

Griffin, A. S. 2004. Social learning about predators: a review and prospectus. Learning
and Behavior 32: 131-140.

Gurevitch, J., and L. V. Hedges. 1999. Statistical issues in ecological meta-analyses.
Ecology 80: 1142-1149.

Hairston, N. G. Jr., W. Lampert, C. E. Céceres, C. L. Holtmeier, L. J. Weider, U. Gaedke,
J. M. Fischer, J. A. F ox, and D. M. Post. 1999. Lake ecosystems - rapid evolution
revealed by dormant eggs. Nature 401: 446.

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

Halvorsen, G., B. K. Dervo, and K. Papinska. 2004. Zooplankton in Lake Atnsjcen 1985-
1997. Hydrobiologia 521: 149-175.

Hanasato, Y. and M. Yasuno. 1985. Effect of temperature in the laboratory studies on
growth egg development and first parturition of five species of Cladocera.
Japanese Journal of Limnology 46: 185-191.

Harvell, C. D. 1990. The ecology and evolution of inducible defenses. Quarterly Review
of Biology 65: 323-340.

Hassel, M. P. 1978. The dynamics of arthropod predator-prey systems. Princeton
University Press, Princeton, New Jersey.

Havel, J. E., and J. B. Shurin. 2004. Mechanisms, effects, and scales of dispersal in
freshwater zooplankton. Limnology and Oceanography 49: 1229-1238.

148

Herzig, A., and B. Auer. 1990. The feeding behavior of Leptodora kindtii and its impact
on the zooplankton community of Neusiedler See (Austria). Hydrobiologia 198:
107-117.

Heath, MR. 1993. The role of escape reaction in determining the size distribution of prey
captured by herring larvae. Environmental Biology of Fishes 38: 331-344.

Heinrich, J. W. 1981. Culture, feeding, and grth of alewives hatched in the laboratory.
Progressive F ish-Culturist 43: 3-7.

Hewett, S. W., and D. J. Stewart. 1989. Zooplanktivory by alevvifes in Lake Michigan:
ontogenetic, seasonal, and historical patterns. Transactions of the American
Fisheries Society 118: 581-596.

Hochberg M. E., and N. J. Gotelli. 2005. An invasions special issue. Trends in Ecology
and Evolution 20: 211.

Hoffman, J. C., M. E. Smith, and J. T. Lehman. 2001. Perch or plankton: top-down
control of Daphnia by yellow perch (Percaflavescens) or Bythotrephes
cederstroemi in an inland lake? Freshwater Biology 46: 759-775.

HOlker, F., H. DOmer, T. Schulze, S. S. Haertel-Borer, S. D. Peacor, and T. Mehner.
2007. Species-specific responses of planktivorous fish to the introduction of a
new piscivore: implications for prey fitness. Freshwater Biology 52: 1793-1806

HOOk, T. O., E. S. Rutherford, D. M. Mason, and G. S. Carter. 2007. Hatch dates, growth,
survival, and overwinter mortality of age-0 alewives in Lake Michigan:
implications for habitat-Specific recruitment success. Transactions of the
American Fisheries Society 136: 1298-1312.

Hovius, J. T., B. E. Beisner, K. S. McCann, and N. D. Yan. 2007. Indirect food web
effects of Bythotrephes invasion: responses by the rotifer Conochilus in Harp
Lake, Canada. Biological Invasions 9: 233-243.

Interlandi, S. J ., and S. S. Kilham. 2001. Limiting resources and the regulation of
diversity in phytoplankton communities. Ecology 82: 1270-1282.

Ives A. R., and A. P. Dobson. 1987. Antipredator behavior and the population dynamics
of simple predator-prey systems. American Naturalist 130: 431-447.

Jackson, J. B. C., M. X. Kirby, W. H. Berger, et al. 2001. Historical overfishing and the
recent collapse of coastal ecosystems. Science 293: 629-638.

Jansen, W. A., and W. C. Mackay. 1992. Foraging in yellow perch, Percaflavescens:
biological and physical factors affecting diel periodicity in feeding, consumption,
and movement. Environmental Biology of Fishes 34: 287-303.

149

Kerfoot, W.C., and L. J. Weider. 2004. Experimental paleoecology (resurrection
ecology): Chasing Van Valen’s Red Queen hypothesis. Limnology and
Oceanography 49: 1300-1 3 1 6.

Kessler, K., and W. Lampert. 2004. Fitness optimization of Daphnia in a trade-off
between food and temperature. Oecologia 140: 381-387.

Kicrboe, T., E. Saiz, and A. Visser. 1999. Hydromechanic signal perception in the
copepod Acartia tonsa. Marine Ecology Progressive Series 179: 97-111.

Kierboe, T., and A. W. Visser. 1999. Predator and prey perception in copepods due to
hydromechanical Signals. Marine Ecology Progressive Series 179: 81-85.

Kuehl, R. O. 2000. Design of experiments: statistical principles of research design and
analysis. 2nd edition. Duxbury Press, California.

Kuhlmann, H. W., J. Kusch, and K. Heckmann. 1999. Predator-induced defenses in
ciliated protozoa. Pages 142—159 in R. Tollrian and C. D. Harvell, editors. The
ecology and evolution of inducible defenses. Princeton University Press,
Princeton, New Jersey, USA.

Laforsch, C., and R. Tollrian. 2004. Inducible defenses in multipredator environments:
Cyclomorphosis in Daphnia cucullata. Ecology 85: 2302-231 1.

Laforsch, C., L. Beccara, and R. Tollrian. 2006. Inducible defenses: The relevance of
chemical alarm cues in Daphnia. Limnology and Oceanography 51: 1466-1472.

Lampert W., and I. Trubetskova. 1996. Juvenile growth rate as a measure of fitness in
Daphnia. Functional Ecology 10: 631-635.

Lantry, B. F ., and R. O'Gorman. 2007. Drying temperature effects on fish dry mass
measurements. Journal of the Great Lakes Research 33: 606-616.

Lass S., and P. Spaak. 2003. Chemically induced anti-predator defenses in plankton: a
review. Hydrobiologia 491: 221-239.

Leech, D. M., A. Padeletti, and C. E. Williamson. 2005. Zooplankton behavioral
responses to solar UV radiation vary within and among lakes. Journal of Plankton
Research 27: 461-471.

Lehman, J. T. 1991. Causes and consequences of cladoceran dynamics in Lake Michigan:
Implications of species invasion by Bythotrephes. Journal Of Great Lakes
Research 17: 437-445.

Lehman, J. T., and D. K. Branstrator. 1995. A model for growth, development, and diet

selection by the invertebrate predator Bythotrephes cederstroemi. Journal of Great
Lakes Research 21: 610-619.

150

Lehman, J. T. and C. E. Caceres. 1993. Food-webs response to species invasion by a
predatory invertebrate: Bythotrephes in Lake Michigan. Limnology and
Oceanography 38: 879-891.

Lehman J. T., D. M. Bilkovic, and C. Sullivan. 1997. Predicting development,
metabolism and secondary production for the invertebrate predator Bythotrephes.
Freshwater Biology 38: 343-352.

Letcher. B. H., J. A. Rice, L. B. Crowder, and K. A. Rose. 1996. Variability in survival of
larval fish: disentangling components with a generalized individual-based model.
Canadian Journal of Fisheries Aquatic Sciences 53: 787-801.

Levin, S. A. 1992. The problem of pattern and scale in ecology. Ecology 73: 1943-1967.

Lima, S. L. 1998. Non-consumptive effects in the ecology of predator-prey interactions.
Bioscience 48: 25-34.

Loose, C. J ., and P. Dawidowicz. 1994. Trade—offs in diel vertical migration by
zooplankton — the costs of predator avoidance. Ecology 75: 2255-2263.

Loose, C. J ., E. von Elert, and P. Dawidowicz. 1993. Chemically-induced diel vertical

migration in Daphnia - A new bioassay for kairomones exuded by fish. Archiv fiir
Hydrobiologie 126: 329-337.

Lohrenz, S. E., G. L. F ahnenstiel, D. F. Millie, O. M. E. Schofield, T. Johengen, and T.
Bergmann. 2004. Spring phytoplankton photosynthesis, growth, and primary
production and relationships to a recurrent coastal sediment plume and river

inputs in southeastern Lake Michigan. Journal of Geophysical Research 109: 1-
13.

Lockwood, J. R. 1998. On the statistical analysis of multiple-choice feeding preferences
experiments. Oecologia 1 16: 475-481.

Lurling M. 2001. Grazing-associated infochemicals induce colony formation in the green
alga Scenedesmus. Protist 152: 7-16.

Luttbeg, B., and O. J. Schmitz. 2000. Predator and prey models with flexible individual
behavior and imperfect information. American Naturalist 155: 669-683.

MacArthur, R. 1955. Fluctuations of animal populations and a measure of community
stability. Ecology 36: 533-536.

MacIsaac, H. J., H. A. M. Ketelaars, I. A. Grigorovich, C. W. Ramcharan, and N. D. Yan.
2000. Modeling Bythotrephes longimanus invasions in the Great Lakes basin
based on its European distribution. Archiv fur Hydrobiologie 149: 1-21.

151

Mack, R. N., D. Simberloff, W. M. Lonsdale, H. Evans, M. Clout, and F. A. Bazzaz.
2000. Biotic invasions: Causes, epidemiology, global consequences, and control.
Ecological Applications 10: 689-710.

Madenjian, C. P., G. L. Fahnenstiel, T. H. Johengen, T. F. Nalepa, H. A. Vanderploeg, G.
W. Fleischer, P. J. Schneeberger, D. M. Benjamin, B. B. Smith, J. R. Bence, E. S.
Rutherford, D. S. Lavis, D. M. Robertson, D. J. Jude, and M. P. Ebner. 2002.
Dynamics of Lake Michigan food web, 1970-2000. Canadian Journal of Fisheries
and Aquatic Sciences 59: 736-753. '

Makarewicz, J. C., P. Bertram, T. Lewis, and E. H. Brown. 1995. A decade of predatory

control of zooplankton species composition of Lake Michigan. Journal of Great
Lakes Research 21: 620-640.

Makarewicz, J. C., P. Bertram, and T. W. Lewis. 1998. Changes in phytoplankton size-
class abundance and species composition coinciding with changes in water
chemistry and zooplankton community structure of Lake Michigan, 1983 to 1992.
Journal of Great Lakes Research 24: 637-657.

Manca, M., C. Ramoni, and P. Comoli. 2000. The decline of Daphnia hyalina in Lago
Maggiore: a comparison of the population dynamics before and after
Oligotrophication. Aquatic Sciences 62: 142-153.

Marsden, J. E., and S. R. Robillard. 2004. Decline of yellow perch in southwestern Lake
Michigan, 1987-1997. North American Journal of Fisheries Management 24: 952-
966.

May, R. M. 1973. Stability and complexity in model ecosystems. Princeton University
Press, Princeton, New Jersey, USA.

McCann, K. The diversity-stability debate. 2000. Nature. 405: 228-233.

McIntosh, A. R., and C. R. Townsend. 1994. Interpopulation variation in mayfly
antipredator tactics - Differential effects of contrasting predatory fish. Ecology
75: 2078-2090.

McIntosh, A. R., and C. R. Townsend. 1995. Impacts of an introduced predatory fish on
mayfly grazing in New Zealand streams. Limnology and Oceanography 40: 1508-
1512.

McNaught, A. S., R. L. Kiesling, and A. Ghadouani. 2004. Changes to zooplankton
community structure following colonization of a small lake by Leptodora kindtii.
Limnology and Oceanography 49: 1239-1249.

Meyer, S. L. 1975. Data analysis for scientists and engineers. Wiley and Sons, Inc.,
Hoboken, NJ.

152

Miller, T., L. B. Crowder, and F. P. Binkowski. 1990. Effects Of changes in the
zooplankton assemblage on growth of bloater and implications for recruitment
success. Transactions of the American Fisheries Society 119: 483—491.

Miller, T. J., L. B. Crowder, and J. A. Rice. 1993. Ontogenetic changes in behavioural
and histological measures of visual acuity in three Species of fish. Environmental
Biology Of Fishes 37: 1-8.

Mills, E. L., J. L. Confer, and R. C. Ready. 1984. Prey selection by young yellow perch:
the influence of capture success, visual acuity, and prey choice. Transactions of
the American Fisheries Society 113: 579-587.

Mills, E. L., et al. 2003. Lake Ontario: food web dynamics in a changing ecosystem
(1970-2000). Canadian Journal of Fisheries and Aquatic Sciences 60: 471-490.

Muirhead, J. R., and H. J. MacIsaac. 2005. Development of inland lakes as hubs in an
invasion network. Journal of Applied Ecology 42: 80-90.

Muirhead, J., and W. G. Sprules. 2003. Reaction distance of Bythotrephes longimanus,
encounter rate and index of prey risk for Harp Lake, Ontario. Freshwater Biology
48: 135-146.

Murphy, T. P., et al. 2003. New microcystin concerns in the lower great lakes. Water
Quality Research Journal of Canada 38: 127-140.

Nelson, E. H., C. E. Matthews, and J. A. Rosenheim. 2004. Predators reduce prey
populations growth by inducing changes in prey behavior. Ecology 85: 1853-
1858.

Neilson, M. A., D. S. Painter, G. Warren, R. A. Hites, I.Basu, D. V. C. Weseloh, D. M.
Whittle, G. Christie, R. Barbiero, M. Tuchman, O. E. Johannsson, T. F. Nalepa,
T. A. Edsall, G. Fleischer, C. Bronte, S. B. Smith, and P. C. Baumann. 2003.
Ecological monitoring for assessing the state of the nearshore and Open waters of
the Great Lakes. Environmental Monitoring and Assessment 88: 103-117.

O'Gorman, R., O. E. Johannsson, and C. P. Schneider. 1997. Age and growth of alewives
in the changing pelagia of Lake Ontario, 1978-1992. Transactions of the
American Fisheries Society 126: 112-126.

Ojaveer, H., M. Simm, and A. Lankov. 2004. Population dynamics and ecological impact
of the non-indigenous C ercopagis pengoi in the Gulf Of Riga (Baltic Sea).
Hydrobiologia 522: 261-269.

Ohman, M. D. 1990. The demographic benefits of diel vertical migration by zooplankton.
Ecological Monographs 60: 257-281.

153

Orcutt, J. D., and K. G. Porter. 1983. Diel vertical migration by zooplankton: Constant
and fluctuating temperature effects on life history parameters of Daphnia.
Limnology and Oceanography 28: 720-730.

Packard, A. T. 2001. Clearance rates and prey selectivity of the predaceous cladoceran
Polyphemus pediculus. Hydrobiologia 442: 177-184.

Paine, R. T. 1980. Food webs: linkage, interaction strength and community infrastructure.
Journal of Animal Ecology 49: 667-685.

Palheimo, J. E. 1974. Calculations of instantaneous birth rate. Limnology and
Oceanography 19: 692-694.

Palmer, A., H. B. Stich, and G. Maier. 2001. Distribution patterns and predation risk of

the coexisting cladocerans Bythotrephes longimanus and Leptodora kindtii in a
large lake - Lake Constance. Hydrobiologia 442: 301-307.

Parej ko, K., and S. I. Dodson. 1991. The evolutionary ecology of an antipredator reaction
norm: Daphnia pulex and Chaoborus americanus. Evolution 45: 1665-1674.

Park, S., S. Chandra, D. C. Muller-Navarra, and C. R. Goldman. 2004. Diel and vertical
variability of seston food quality and quantity in a small subalpine
Oligomesotrophic lake. Journal of Plankton Research 26: 1489-1498.

Pastorok, R. A. 1981. Prey vulnerability and size selection by Chaoborus larvae. Ecology
62:1311-1324.

Peacor, S. D. 2001. The contribution of trait-mediated indirect effects to the net effects of
a predator. Proceedings of the National Academy of Science 98: 3904-3908.

Peacor, S. D. 2002. Positive effect of predators on prey grth rate through induced
modifications of prey behaviour. Ecology Letters 5: 77-85.

Peacor, S. D., and E. E. Werner. 2004. How dependent are species-pair interactions
strengths on other species in the food web? Ecology 85: 2754-2763.

Pearl, C. A., M. J. Adams, G. S. Schuytema, and A. V. Nebeker. 2003. Behavioral
responses of anuran larvae to chemical cues of native and introduced predators in

the Pacific Northwestern United States. Journal of Herpetology 37: 572-576.

Peckarsky, B. L., C. A. Cowan, M. A. Penton, and C. Anderson. 1993. Sublethal
consequences of stream-dwelling predatory stoneflies on mayfly growth and
fecundity. Ecology 74: 1836-1846.

Pichlova, R., and Z. Brandl. 2003. Predatory impact of Leptodora kindtii on zooplankton
community in the Slapy Reservoir. Hydrobiologia 504: 177-184.

154

Post, J. R. 1990. Metabolic allometry of larval and juvenile yellow perch (Perca
flavescens): in situ estimates and bioenergetics models. Canadian Journal of
Fisheries and Aquatic Science 47: 554-560.

Pothoven, S. A., G. L. F ahnenstiel, and H. A. Vanderploeg. 2003. Population dynamics
of Bythotrephes cederstroemii in south-east Lake Michigan 1995-1998.
Freshwater Biology 46: 1491-1501.

Pothoven, S. A., H. A. Vanderploeg, J. F. Cavaletto, D. M. Krueger, D. M. Mason, and S.
B. Brandt. 2007. Alewife planktivory control the abundance of two invasive
predatory cladocerans in Lake Michigan. Freshwater Biology 52: 561-573.

Ramcharan, C. W., and W. G. Sprules. 1986. Visual predation in Mysis relicta Loven.
Limnology and Oceanography 31: 414-420.

Reichwaldt, E. S., I. D. Wolf, and H. Stibor. 2004. Effects of a fluctuating temperature
regime experienced by Daphnia during diel vertical migration on Daphnia life
history parameters. Hydrobiologia 543: 199-205.

Riessen H. P., J. W. Sommerville, C. Chiappari and D. Gustafson. 1988. Chaoborus
predation, prey vulnerability, and their effect in zooplankton communities.
Canadian Journal of Fisheries and Aquatic Sciences 45: 1912-1920.

Rivier, I. K. 1998. The predatory cladocera (Onychopoda: Podonidae, Polyphemoidae,
Cercopagidae) and Leptodoridae of the world. Backhuys Publishers, Leiden, NL.

Roche, K. 1990. Some aspects of vulnerability to cyclopoid predation on zooplankton
prey individuals. Hydrobiologia 198: 153-162.

Roff, J. C., and R. R. Hopcroft. 1986. High precision microcomputer based measuring
system for ecological research. Canadian Journal of Fisheries and Aquatic
Sciences 43: 2044-2048.

Rosa, F ., and N. M. Burns. 1987. Lake Erie central basin oxygen depletion changes from
1929-1980. Journal of Great Lakes Research 13: 684-696.

Rose, K. A., and J. H. Cowan. 1993. Individual-based model of young-of-the-year striped
bass population dynamics. 1. Model description and baseline simulations.
Transactions of the American Fisheries Society 122: 415-438.

Rosenzweig, M. 1971. The Paradox of Enrichment. Science 171: 385-387.

Rothhaupt, K. O. 1991. Variations on the zooplankton menu: A reply to the comment by
Smith. Limnology and Oceanography 36: 824-827.

Scavia, D., Fahnenstiel, G.L., Evans, M.S., Jude, D.J., and Lehman, J .T. 1986. Influence
of salmonid predation and weather on long-terrn water quality trends in Lake
Michigan. Canadian Journal of Fisheries and Aquatic Sciences 43: 435—443.

155

Schmitz, O. J. 2005. Scaling from plot experiments to landscapes: studying grasshoppers
to inform forest ecosystem management. Oecologia 145: 225-234.

Schmitz, O. J ., V. Krivan, and O. Ovadia. 2004. Trophic cascades: the primacy of trait-
mediated indirect interactions. Ecology Letters 7: 153-163.

Schulz, K. L., and P. M. Yurista. 1999. Implications of an invertebrate predator's
(Bythotrephes cederstroemi) atypical effects on a pelagic zooplankton
community. Hydrobiologia 380: 179-193.

Schulze, P. C., and A. S. Brooks. 1987. The possibility of predator avoidance by Lake
Michigan zooplankton. Hydrobiologia 146: 47-56.

Shave C. R., C. R. Townsend, and T. A. Crow]. 1994. Antipredator behaviors of a fresh-
water crayfish (Paranephrops zealandicus) to a native and introduced predator.
New Zealand Journal of Ecology 18: 1-10.

Shuter, B. J., and K. K. Ing. 1997. Factors affecting the production of zooplankton in
lakes. Canadian Journal of Fisheries and Aquatic Sciences 54: 359-377.

Shuter, B. J., and D. M. Mason. 2001. Exotic invertebrates, food-web disruption, and lost
production: understanding impacts of dreissenid and cladoceran invaders on
lower-lakes fish communities and forecasting invasion impact on upper-lakes fish
communities. Prepared for Board of Technical Experts, Great Lakes Fishery
Commission.

Simberloff, D. 1995. Why do introduced Species appear to devastate islands more than
mainland areas? Pacific Science. 49: 87-97.

Semes, T. A., and D. L. Aksnes. 2006. Concurrent temporal patterns in light absorbance
and fish abundance. Marine Ecology Progressive Series 325: 181-186.

Stabell, O. B., F. Ogbebo, and R. Primicerio. 2003. Inducible defences in Daphnia
depend on latent alarm signals from conspecific prey activated in predators.
Chemical Senses 28: 141-153.

Stemberger, R. S., and J. J. Gilbert. 1987. Multiple-species induction of morphological
defenses in the rotifer Keratella testudo. Ecology 68: 370—378.

Sterner, R. W., and M. S. Schwalbach. 2001. Diel integration of food quality by Daphnia:
Luxury consumption by a fi-eshwater planktonic herbivore. Limnology and
Oceanography 46: 410-416.

Stewart, D. J., and F. P. Binkowski. 1986. Dynamics of consumption and food

conversion by Lake Michigan alewives: an energetic-modeling synthesis.
Transactions of the American Fisheries Society 115: 643-661.

156

Stich, H., and W. Lampert. 1981. Predator evasion as an explanation of diurnal vertical
migration by zooplankton. Nature 293: 396-398.

Stich, H. B., and W. Lampert. 1984. Growth and reproduction of migrating and non-
migrating Daphnia species under simulated food and temperature conditions of
diurnal vertical migration. Oecologia 61: 192—196.

Straile, D., and A. Halbich. 2000. Life history and multiple antipredator defenses of an
invertebrate pelagic predator, Bythotrephes longimanus. Ecology 81: 150-163.

Strauss, S. Y., J. A. Lau, and S. P. Carroll. 2006. Evolutionary response of natives to
introduced species: what do introductions tell us about natural communities?
Ecology Letters 9: 357-374.

Strayer, D. L., V. T. Eviner, J. M. Jeschke, and M. L. Pace. 2006. Understanding the
long-terms effects of species invasion. Trends in Ecology and Evolution 21: 645-
651.

Strecker, A. L., S. E. Amott, N. D. Yan, and R. Girard. 2006. Variation in the response of
crustacean zooplankton species richness and composition to the invasive predator

Bythotrephes longimanus. Canadian Journal of Fisheries and Aquatic Sciences 63:
2126-2136.

Tanner, C. J., and D. K. Branstrator. 2006. Generational and dual-species exposures to
invertebrate predators influence relative head size in Daphnia mendotae. Journal
of Plankton Research 28: 793-802.

Tiselius, P., and P. R. Jonsson. 1990. Foraging behavior or 6 calanoid copepods —
observations and hydrodynamic analysis. Marine Ecology Progressive Series 66:
23-33.

Tollrian, R. 1995. Predator-induced morphological defenses: Costs, life history shifts,
and maternal effects in Daphnia pulex. Ecology 76: 1691-1705.

Tollrian, R., and C. D. Harvell [eds.]. 1999. The ecology and evolution of inducible
defenses. Princeton University Press, New Jersey, USA.

Trussell, G. C. 2000. Predator-induced morphological trade-offs in latitudinally-separated
populations of Littorina obtusata. Evolutionary Ecology Research 2:803-822.

Turner, A. M., and G. G. Mittelbach. 1990. Predator avoidance and community structure

- interactions among piscivores, planktivores, and plankton. Ecology 71: 2241-
2254.

Vanderrneer, J. 1969. The competitive structure of communities: an experimental
approach with Protozoa. Ecology 50: 362-372.

157

Vanderploeg, H. A., J. A. Bowers, O. Chapelski, and H. K. 800. 1982. Measuring in situ
predation by Mysis relicta and observations on underdispersed microdistribution
of zooplankton. Hydrobiologia 93: 109-119.

Vanderploeg, H. A., J. R. Liebig, and M. Omair. 1993. Bythotrephes predation on Great-
Lakes zooplankton measured by an in situ method - implications for zooplankton
community structure. Archiv fiir Hydrobiologie 127: 1-8.

Vanderploeg H. A., T. F. Nalepa, D. J. Jude, E. L. Mills, K. T. Holeck, J. R. Liebig, I. A.
Grigorovich, and H. Ojaveer. 2002. Dispersal and emerging ecological impacts of

Ponto-Caspian species in the Laurentian Great Lakes. Canadian Journal of
Fisheries and Aquatic Sciences 59: 1209-1228.

Viitasalo, M., T. Kicrboe, J. Flinkman, L. W. Pedersen, and A. W. Visser. 1998.
Predation vulnerability of planktonic copepods: consequences of predator

foraging strategies and prey sensory abilities. Marine Ecology Progressive Series
175: 129-142.

Walters, C., D. Pauly, V. Christensen 1999. Ecospace: Prediction of mesoscale spatial
patterns in trophic relationships of exploited ecosystems, with emphasis on the
impacts of marine protected areas. Ecosystems 2: 539—554

Walton, W. E., N. G. Hairston, and J. K. Wetterer. 1992. Growth-related constraints on
diet selection by sunfish. Ecology 73: 429-437.

Weider, L. J ., and W. Lampert. 1985. Differential response of Daphnia genotypes to
oxygen stress: respiration rates, hemoglobin content and low-oxygen tolerance.
Oecologia 65: 487-491.

Werner, E. E., J. F. Gilliam, D. J. Hall, and G. G. Mittelbach. 1983. An experimental test
of the effects of predation risk on habitat use in fish. Ecology 64: 1540-1548.

Werner, E. E., and S. D. Peacor. 2003. A review of trait-mediated indirect interactions in
ecological communities. Ecology 84: 1083-1100.

Wetzel, R. G. 2001. Limnology: lake and river ecosystems. 3rd edition. Academic Press,
San Diego, California.

Wilder S. M., and A. L. Rypstra. 2004. Chemical cues from an introduced predator
(Mantodea, Mantidae) reduce the movement and foraging of a native wolf spider

(Aranae, Lycosidae) in the laboratory. Environmental Entomology 33: 1032-
1036.

Williamson, C. E., and R. E. Magnien. 1981. Diel vertical migration in Mesocyclops
edax: implications for predation rate estimates. Journal of Plankton Research 4:
329-339.

158

Williamson, C. E., R. W. Sanders, R. E. Moeller, and P. L. Stutzman. 1996. Utilization of
subsurface food resources for zooplankton reproduction: implications for diel
vertical migration theory. Limnology and Oceanography 41: 224—233.

Winder, M., M. Boersma, and P. Spaak. 2003. On the cost of vertical migration: are

feeding conditions really worse at greater depths? Freshwater Biology 48: 383-
393.

Wissel, B., W. J. Boeing, and C. W. Ramcharan. 2003. Effects of water color on predator
regimes and zooplankton assemblages in freshwater lakes. Limnology and
Oceanography 48: 1965-1976.

Wojtal, A., P. Frankiewicz, I. Wagner-Lotkowska, and M. Zalewski. 2004. The
evaluation of the role of pelagic invertebrate versus vertebrate predators on the

seasonal dynamics of filtering Cladocera in a shallow, eutrophic reservoir.
Hydrobiologia 515: 123-135.

Wootton, J. T. 1994. Putting the pieces together - testing the independence of interactions
among organisms. Ecology 75: 1554-1551.

Wright, D., and J. Shapiro. 1990. Refuge availability: a key to understand the summer
disappearance of Daphnia. Freshwater Biology 24: 43-62.

Yan N. D., A. Blukacz, W. G. Sprules, P. K. Kindy, D. Hackett, R. E. Girard, and B. J.
Clark. 2001. Changes in zooplankton and the phenology of the Spiny water flea,
Bythotrephes, following its invasion of Harp Lake, Ontario, Canada. Canadian
Journal of Fisheries and Aquatic Sciences 58: 2341-2350.

Yan, N. D., R. Girard, and S. Boudreau. 2002. An introduced invertebrate predator
(Bythotrephes) reduces zooplankton species richness. Ecology Letters 5: 481-485.

Yan, ND. and T.W. Pawson 1998. Seasonal variation in the size and abundance of the
invading Bythotrephes in Harp Lake, Ontario, Canada. Hydrobiologia 361: 157-
168.

Yurista, P. M., and K. L. Schulz. 1995. Bioenergetic analysis of prey consumption by
Bythotrephes cederstroemi in Lake Michigan. Canadian Journal of Fisheries and
Aquatic Sciences 52: 141-150.

Young, S., and V. A. Taylor. 1987. Spontaneous and evoked eye-movements in
Polyphemus pediculus (Cladocera, Crustacea) - a case of open-loop tracking.
Journal of Experimental Biology 131: 323-336.

Zar J. H. 1999. Biostatistical analysis, 4th ed. Prentice-Hall, Inc., Englewood Cliffs, N. J.

159

lllllllfllllllllljljll[1318111l