 

This is to certify that the
dissertation entitled

Regional coexistence and local dominance in Chaoborus:
species sorting along a predation gradient

presented by

Erica Ann Garcia

has been accepted towards fulfillment
of the requirements for the

PhD. degree in Zoology and Program in
Ecology, Evolutionary Biology
and Behavior

 

 

  

(”f/X’ﬁz

’ v ’ Major Professor’s Signature

8/Z/Qé;

Date

MSU is an Afﬁrmative Action/Equal Opportunity Institution

 

LIBRARY
Michigan State

 

 

 

 

University

 

4 _ —‘-.—.—--.-.--l—l—.- —.-.-.—.-.-.—.-.-.-.-.-.-.—.—-—-.---—¢-«-.- 4

 

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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/ClRC/DateDue.p65.p. 15

 

 

REGIONAL COEXISTENCE AND LOCAL DOMINANCE IN CHAOBORUS:
SPECIES SORTING ALONG A PREDATION GRADIENT
By

Erica Ann Garcia

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Zoology
Program in Ecology, Evolutionary Biology, and Behavior

2006

ABSTRACT

REGIONAL COEXISTENCE AND LOCAL DOMINANCE IN CHAOBORUS:
SPECIES SORTING ALONG A PREDATION GRADIENT
By

Erica Ann Garcia

The non-random distributional patterns of species across the landscape may be a
function of dispersal and local Species sorting based on the traits of the species involved.
In aquatic systems in particular, the direct and indirect effects of predators are major
factors limiting species distributions and abundances. However, while a number of
taxonomic groups have been shown to vary in their distribution based on the presence or
absence of ﬁsh, far less is known about the distributional patterns of species across a
landscape that varies in the intensity of ﬁsh predation (due either to variation in ﬁsh
density or the degree to which organisms have a refuge from ﬁsh). In my dissertation
research, I used four species Of phantom midge larvae (Chaoborus) found in
southwestern Michigan, to examine how different species’ traits are functionally related
to species sorting along natural and experimental gradients in predation.

Through ﬁeld surveys of 17 lakes and ponds, a long term data set in one lake and
empirical work in cattle tanks and replicate ponds at the Experimental Pond Facility at
the Kellogg Biological Station, I found that species traits such as pigmentation, size,
oviposition habitat selection (OHS) and diel vertical migration (DVM) are key in how
Chaoborus species sort across the landscape. This is due to the fact that the four
Chaoborus species in this study varied relative to these traits and I found that this
variation was related to differences in vulnerability to ﬁsh. Thus, C haoborus species’ trait
variation together with variation across the landscape in ﬁsh predation intensity lead to

the Observed patterns of C haoborus Species’ regional coexistence and local dominance.

Overall, this research highlights the importance of considering environmental gradients
(versus simply presence or absence of a factor) when examining the distribution and

abundance of organisms.

F or my parents, Guillermo and Vicki Garcia

And for A buelita, te quiero mucho

iv

ACKNOWLEDGMENTS

This research couldn’t have made it from start to ﬁnish if not for the support and
advice of many friends and colleagues. My ﬁrst big thank you goes to my advisor, Gary
Mittelbach. Gary has been great, his encouragement, assistance and advice were always
there when I most needed it. I am very grateful for his support over the past 6 years. I
also want to thank the rest of the faculty that served on my committee during various
stages of this dissertation: Kay Gross, Scott Peacor, Ace Samelle, and Alan Tessier. In
particular, I appreciated Kay inviting me to be an honorary member of her lab when my
own lab had only 2 members.

My transition to life in rural Michigan and graduate school was made possible
by the wonderful community of people here at KBS, the beautiful fall colors and those
spectacular winter days when there is not a cloud in the sky and the ground is completely
covered in white.

My lab mates Tara Darcy-Hall, Nate Dom, Chris Steiner and Jeremy Wojdak
helped get me started and kept me going. Chris showed me the ways of the zooplankton
and Nate the ways of ﬁsh, I am indebted to them both for their mentorship, friendship and
help in the ﬁeld. I also appreciated the moral and intellectual support of Tara and Jeremy.
I couldn’t have imagined life in the Mittelbach lab without them. And then there was
one. . .and I gained new lab mates, Sarah Emery, Greg Houseman and Rich Smith, thanks
for taking me in.

My cohort —- Meg Duffy, Sarah Emery, Angie Roles, Heather Sahli, and honorary
member Frances Knapczyk, what can I say but you have been awesome! I survived
graduate school in large part because of your friendship and support. My soccer mates
and our few fans throughout the years have provided a much needed outlet and are
another part of the reason I made it to this point. The best house mate a person could ask

for, Pamela Geddes was always enthusiastic, kind and a true friend.

There are also many others I need to thank for assistance in the ﬁeld and lab.
Most especially, Tara Darcy-Hall, Nate Dorn, Meg Duffy, Jeremie Fant, Mathew
Leibold, Gary Mittelbach, Chris Steiner, Becky Tonietto, and Pam Woodruff. I also
would like to thank the amazing support staff here at KBS, Char Adams, Stu Bassett,
Nina Consolatti, Alice Gillespie, John Gorentz, Janell Lovan, Mike Martin, Sally Shaw
and Melissa Yost.

I also must acknowledge the incredible science mentors who introduced me to the
wonders of nature and scientiﬁc exploration and set me on this path, Dee Strange, Linda
Dillard and Harry Greene.

Last but not least I need to thank the other good friends I made along the way and
my family. I really enjoyed my conversations and walks with Heather Wojdak. Jeremie
Fant and Jon Clark have always been willing providers of much love and support, in good
times and bad. My parents, Vicki and Guillermo and my brother Chris have always
believed in me and I am so very grateful for their unconditional love. I can’t thank them

all enough.

vi

TABLE OF CONTENTS

LIST OF TABLES ........................................................................................................ ix
LIST OF FIGURES ....................................................................................................... x
CHAPTER 1
INTRODUCTION ......................................................................................................... l
C haoborus Background ................................................................................................. 2
Dissertation Synopsis .................................................................................................... 4
CHAPTER 2
CHA 080R US SPECIES DISTRIBUTION AND ABUNDANCE ACROSS
NATURAL GRADIENTS TN PLANKTIVORY .......................................................... 7
Abstract .......................................................................................................................... 7
Introduction ................................................................................................................... 7
Methods ......................................................................................................................... 9
Field survey ....................................................................................................... 9
Statistical analyses ........................................................................................... 13
Results ......................................................................................................................... 13
Discussion .................................................................................................................... 19
CHAPTER 3
OVIPOSITION HABITAT SELECTION AND LARVAL PERFORMANCE ......... 24
Abstract ........................................................................................................................ 24
Introduction ................................................................................................................. 24
Methods ....................................................................................................................... 26
Statistical analyses ........................................................................................... 29
Results ......................................................................................................................... 29
Mesocosm experiment ..................................................................................... 29
Bag experiment ................................................................................................ 32
Discussion .................................................................................................................... 32
CHAPTER 4
CHAOBOR US SPECIES SORTING ALONG A PREDATION GRADIENT ........... 36
Abstract ........................................................................................................................ 36
Introduction ................................................................................................................. 36
Methods ....................................................................................................................... 38
Study organisms .............................................................................................. 38
Fish gradient experiments ................................................................................ 38
Statistical analyses ........................................................................................... 41
Prey preference trials: non-refuge and refuge ................................................. 41
Results ......................................................................................................................... 43
Fish gradient experiments: Species sorting and colonization .......................... 43
Prey preference trials: non-refuge and refuge ................................................. 51
Discussion .................................................................................................................... 54

vii

LITERATURE CITED ................................................................................................ 6O

viii

LIST OF TABLES
CHAPTER 1
Table 1.1. C haoborus species studied and their associated traits ................................. 6
CHAPTER 2

Table 2.1. Means and ranges (in parentheses) of limnological variables measured in
the study ponds and lakes over the survey period ....................................................... 10

CHAPTER 4

Table 4.1. Fish gradient experimental design in A) Experiment 1 and B) Experiment
2, with initial and ﬁnal ﬁsh biomass ............................................................................ 45

ix

LIST OF FIGURES
CHAPTER 1

Figure 1.1. C. americanus. Photograph by C. Steiner .................................................. 3

CHAPTER 2

Figure 2.1. Presence of four Chaoborus species in 17 lakes and ponds that vary in
presence/absence of ﬁsh and depth ............................................................................. 15

Figure 2.2. Seasonal dynamics of Chaoborus densities in ﬁve southwestern Michigan
lakes (mean densities based on three samples per date) .............................................. 16

Figure 2.3. Seasonal dynamics of Chaoborus densities in six southwestern Michigan
ﬁshless ponds (mean densities based on three samples per date) ................................ 17

Figure 2.4. Association of 5 lakes and 6 ponds based on NMDS ordination of
Chaoborus species abundances. The angle and length of the joint plot lines indicate
the direction and strength of the relationship of environmental variables with the
ordination scores. Asterisks represent the centroids for each Chaoborus species ...... 18

Figure 2.5. Chaoborus density in Wintergreen Lake as a function of variation in
planktivore density over 16 yrs. Lines are the ﬁtted linear regressions for each species
(the solid line is signiﬁcant, p<0.01). Planktivore densities determined by
mark/recapture (see Mittelbach et al. 2006 for details) ............................................... 20

CHAPTER 3

Figure 3.1. Effects of ﬁsh cue, ﬁsh, and high productivity on mean Chaoborus
density (number per liter i 1 SE) averaged across all sampling dates ........................ 30

Figure 3.2. C haoborus density (number per liter) in the ﬁshless treatments of the
mesocosm experiment. Each point is a sampling date for an individual tank. (C. other
includes C. punctipennis and C. ﬂavicans) .................................................................. 31
Figure 3.3. Effects of ﬁsh cue on mean density (number per liter :t 1 SE) of three
Chaoborus species on the ﬁnal day of the bag experiment ......................................... 33

CHAPTER 4

Figure 4.1. Change in Chaoborus density (number per liter) through time in
Experiment 2 grouped into three ﬁsh density categories. Time is the sampling day of

the experiment (0-1 12, May-September 2005). The ﬁlled circles are the means (:I: l
s.e., n=4) for the zero ﬁsh biomass ponds, the open circles are the means (i 1 s.e.,

=4) for the medium ﬁsh biomass (500-1250 g) ponds, and the ﬁlled triangles are the
means (i 1 s.e., n=7) for the high ﬁsh biomass (3250-5850 g) ponds ........................ 46

Figure 4.2. Chaoborus density (number per liter) in Experiment 2 averaged across
the last three sampling dates as a function of ﬁnal ﬁsh biomass. Lines are the ﬁtted
linear regressions for each species (solid lines are signiﬁcant, p<0.01) and each point
represents a pond ......................................................................................................... 47

Figure 4.3. Number of days C. americanus was present in the water column in
Experiment 2 as a function of ﬁnal ﬁsh biomass. The solid line is the ﬁtted linear
regression (p<0.01) and each point represents a pond ................................................. 49

Figure 4.4. Change in Chaoborus density (number per liter) through time in
Experiment 1 grouped into three ﬁsh biomass categories. Time is the sampling day of
the experiment (0-123, May-September 2003). The ﬁlled circles are the means (d: 1
s.e., n=3) for the zero ﬁsh biomass ponds, the open circles are the means (i l s.e.,
n=2) for the medium ﬁsh biomass (2300-3200 g) ponds, and the ﬁlled triangles are
the means (i: l s.e., n=5) for the high ﬁsh biomass (5700-12,050 g) ponds ............... 50

Figure 4.5. C haoborus density (number per liter) in Experiment 1 averaged across
the last ﬁve sampling dates as a function of ﬁnal ﬁsh biomass. Lines are the ﬁtted
linear regressions for each species (solid lines are signiﬁcant, p<0.01) and each point
represents a pond ......................................................................................................... 52

Figure 4.6. Average preference (Manly-Chesson a i l s.e.) of Bluegill for three
species of Chaoborus in a) 8 non-refuge feeding trials and b) 4 refuge feeding trials.
The line indicates no preference .................................................................................. 53

Figure 4.7. Frequency distribution of Chaoborus lengths (includes C. punctipennis,
C. ﬂavicans, and C. americanus) found in the environment (gray bars) and ﬁsh diet
(black bars) in a) non-refuge feeding trials (n=8 trials) b) refuge feeding trials (n=4
trials) ............................................................................................................................ 55

xi

CHAPTER 1
INTRODUCTION

Local communities are composed of a subset of the regional species pool that can
disperse to a locality, survive abiotic conditions and persist given the local biotic
interactions (e.g. competition, predation) (Urban 2004). Historically, much of community
ecology theory has focused on local interactions (competition and predation; Tilman
1982, Leibold 1996) independent of regional inﬂuences (colonization and extinction;
MacArthur and Wilson 1967). However, in the last decade, community ecologists have
acknowledged the metacommunity concept as an important way to better understand the
linkages between regional and local factors in structuring communities.

Four paradigms of the metacommunity approach have been described and include
two that assume local sites differ only with respect to the species composition at a given
point in time (patch-dynamic and neutral) and two that assume local sites are
heterogeneous and therefore different species are favored at different locales (species
sorting and mass effects) (Leibold et al. 2004). In a recent meta-analysis, Cottenie (2005)
Observed that habitat heterogeneity and species sorting dynamics were dominant
processes structuring communities. He also found that the type of dispersal (passive
versus active) was an important determinant of metacommunity type and passive
dispersers as well as active dispersers in marine and lake habitats were strongly related to
environmental dynamics.

In my research, I examined the distribution and abundance of active dispersing
lake and pond species in the genus C haoborus (Diptera: Chaoboridae) and used the
Species sorting approach to determine the relative importance of environmental factors
known to vary across the landscape (e. g. ﬁsh planktivory, productivity, pH, etc) in

structuring this species assemblage.

Chaoborus background

The larvae of the phantom midge C haoborus are important members of many
freshwater food webs. They are distributed worldwide with 12 described species in North
America (Swther 1970) and are often the dominant invertebrate predators found in the
plankton of lakes and ponds. C haoborus are omnivorous, gape-limited, ambush predators
of small to medium sized zooplankton (e.g., Moore et al. 1994, Swift and Fedorenko
1975, Pastorok 1981). In north temperate lakes, species of Chaoborus reveal patterns in
distribution across the landscape, strongly suggesting the importance of ﬁsh predation
and possibly other factors such as, interspeciﬁc interactions, water transparency,
temperature, and nutrient levels (Pope et al. 1973, von Ende 1979, Lamontagne et al.
1994, Wissel et al. 2003). Here, I focus on the four Chaoborus species found in

southwestern Michigan: C. americanus, C. punctipenm's, C. ﬂavicans, and C. albatus.

 

Figure 1.1. C. americanus. Photograph by C. Steiner.

Dissertation synopsis

In my dissertation research, I used a combination of ﬁeld surveys of about 20
lakes and ponds, a long term data set in one lake, experimental work in mesocosms and
two large-scale predator manipulations in ponds, to elucidate the processes that explain
the non-random distributional pattern of the Chaoborus species assemblage.

In Chapter 2 I asked: What environmental factors are important to the
distributional pattern of the Chaoborus species assemblage? My approach was to ﬁrst
document the pattern in C haoborus species distribution and abundance in nature, across
aquatic systems that varied in the level of ﬁsh planktivory, productivity, area, depth,
conductivity, pH, temperature and dissolved oxygen. I then compared these results to the
changes in C haoborus community composition in response to a documented decrease and
then increase in planktivore density over 16 years within a single lake. I found that the
level of ﬁsh planktivory, in addition to environmental variables correlated to the level of
ﬁsh planktivory, (i.e. lake area and depth; Tessier and Woodruff 2002) are important
factors related to the distribution and abundance of Chaoborus species across the
landscape.

In Chapter 3 I asked: DO Species of Chaoborus exhibit directed dispersal
(oviposition habitat selection; OHS) and can this explain the presence/absence patterns of
C haoborus species across the landscape? I tested the hypothesis that the Chaoborus
species most vulnerable to predation by ﬁsh (C. americanus) will exhibit OHS but that
species that are less vulnerable to ﬁsh predation (C. punctipennis and C. ﬂavicans) will
not. The results from one experiment supports this hypothesis but in a second experiment
all species of Chaoborus were observed to show no discrimination between sites that
were ﬁshless and those with ﬁsh cues. The discrepancy between these two experiments
may be explained by strong Site ﬁdelity of C. americanus.

In Chapter 4 I asked two questions: 1) How does predator density affect

C haoborus species composition and relative abundance? and 2) Is the pattern of prey

choice by an important planktivorous ﬁsh, the bluegill sunﬁsh (Lepomis macrochirus),
consistent with the observed pattern of Chaoborus species sorting across the predation
gradient? I investigated the ﬁrst question in two pond experiments where I examined
Chaoborus species sorting along an experimental gradient in bluegill density. I also
experimentally tested bluegill prey choice for the four Chaoborus species found in
southwestern, MI. The results of these experiments indicate that 1) the gradient in ﬁsh
density can lead to clear species sorting within the species assemblage that is consistent
with distributional patterns observed in nature, and 2) that larval body size, transparency,
and diel vertical migration (DVM) behavior are important traits determining the

vulnerability of different Chaoborus species to planktivorous ﬁsh such as the bluegill.

Table 1.1. Chaoborus species studiedand their associated traits

 

 

Species 4‘h instar length (mn) Trarsparem OHS DVM References
C. armricanus 10-13 x C3115; 21;:VI’15719977,
C ﬂavicwzs 942.7 x x wiggiif‘oig’twe
C. pmnpemrs 7.5-9.5 x x ngerg
C albums 7-9.4 x X Tjossem 1990

 

Note: References point to selected papers on the change in the inensity of the DVM belavior associated with species
experiencing a new environment (ﬁsh-ﬁshless or ﬁshless- ﬁsh).

CHAPTER 2
CHA OBOR US SPECIES DISTRIBUTION AND ABUNDANCE ACROSS NATURAL
GRADIENTS IN PLANKTIVORY

Abstract

Species traits are functionally related to the determinants of species distributions
and development and maintenance of community structure. Here, 1 present data on the
distribution and abundance of four species of Chaoborus (Diptera: Chaoboridae) that
vary in species traits (i.e. pigmentation, diel vertical migration (DVM) behavior and size)
that are important to coexisting with ﬁsh predators. I examine the spatial and seasonal
pattern of the Chaoborus species assemblage in 1) a ﬁeld survey of 11 lakes and ponds
that vary in environmental factors such as the level of ﬁsh planktivory, pH, productivity,
area and depth, and 2) a long term data set from Wintergreen Lake that experienced large
changes in the levels of ﬁsh planktivory but varied little with respect to other
environmental variables. Chaoborus americanus is the largest species and does not
exhibit DVM behavior, only occurred in ponds without ﬁsh. Chaoborus albatus is
relatively small and does exhibit DVM behavior, was only found in shallow lakes and
ponds with ﬁsh. C haoborus punctipennis is also small and exhibits DVM behavior, was
found in all lakes and ponds with ﬁsh; and C. ﬂavicans is relatively large and exhibits
DVM behavior, was found in both ﬁsh and ﬁshless lakes and ponds. The results of this
study clearly document the importance of ﬁsh predators as drivers of distribution and
abundance in this assemblage.
Introduction

The distribution of species across the landscape may be a function of dispersal
and local species sorting based on the traits of the species involved. The inﬂuence of
particular species traits is especially apparent when a group of closely related species
shows a clear distributional pattern across a strong environmental gradient. The presence

of planktivorous ﬁsh has been considered to be a key factor contributing to the

distribution and abundance of Chaoborus (Diptera: Chaoboridae) across a landscape of
lakes and ponds (e.g. Pope et al. 1973, von Ende 1979, and Wissel et al 2003). The larvae
of the phantom midge Chaoborus are important members of many freshwater food webs
as they are frequently the dominant invertebrate predators in these systems. They are
often the largest organisms in the plankton and are gape-limited, ambush predators of
small to medium sized zooplankton (e.g., Moore et al. 1994, Swift and F edorenko 1975,
Pastorok 1981). Species traits such as small size and vertical migration (DVM) behavior,
where larvae are found in the water column at night and then migrate into the sediments
or deeper water during the day, are considered to be effective strategies that allow some
species to co-occur with ﬁsh (Dawidowicz et al. 1990, Tjossem 1990, Berendonk et al.
2003). Of the 12 described C haoborus species in North America (Seether 1970), four
inhabit lakes and ponds in southwestern Michigan: C. americanus, C. punctipennis, C.
ﬂavicans, and C. albatus. These four species vary widely in morphology from the
relatively large, opaque C. americanus (4th instar length 10-13 mm) to the similar-sized
but transparent C. ﬂavicans (4th instar length 9-12.7 mm) to the small, transparent C.
punctipennis (4th instar length 7.5-9.5 mm) and C. albatus (4th instar length 7-9.4 mm)
(Cook 1956, Swther 1970).

The objective of this study was to document the spatial and seasonal pattern of the
Chaoborus species assemblage in southwestern Michigan and to determine the effect of
environmental variables, particularly the level of ﬁsh planktivory, on species distribution
and abundance. Although other studies have examined patterns of C haoborus distribution
most have only sampled during the day (Pope et al. 1973), or collected data from single
visits to each lake or pond (Wissel et al. 2003) and few have directly looked at the effect
of ﬁsh predation. While ﬁsh planktivory varies across the well-established environmental
gradient in lentic freshwater habitats so do other environmental factors (Wellbom et al.
1996, Tessier and Woodruff 2002). Therefore, I also made use of a unique opportunity to

examine C haoborus abundances in a single lake that varied in level of ﬁsh planktivory

over a 16-year period due to the extinction and purposeful reintroduction of ﬁsh species.
Here, I present data on Chaoborus species presence/absence gathered from multiple
surveys and experiments as well as data on the Chaoborus Species assemblage from a
ﬁeld survey of 11 ponds and lakes that were sampled throughout the growing season of
one year. In addition, I include a long-term data set on ﬁsh and Chaoborus dynamics to
examine how C haoborus community composition responded to a gradual decrease and
then increase in planktivore density over 16 years within a single lake that varied little in
abiotic conditions through time.

Methods

Field Survey

I conducted a ﬁeld survey of 11 lakes and ponds located within a 100 km radius
of the W. K. Kellogg Biological Station (KBS, Hickory Corners, Michigan, USA). All
ponds contained water year-round at least two years prior to the study (personal
observation). These systems were chosen to encompass a gradient in levels of planktivory
(see Tessier and Woodruff 2002) and productivity (total phosphorus). Pond surface area
and depth were measured on 25 June 2004; lake surface area and depth were collected
from the literature (Table 2.1, Tessier and Woodruff 2002, Caceres and Tessier 2004,
Mittelbach et al. 2006).

Ponds and lakes were sampled every 2-4 weeks in 2004, beginning in mid-May
and ending in early September. Chaoborus were collected one hour after sunset by taking
three vertical tows through the entire water column with a 30 cm diameter, 500-micron
plankton net, at the deepest point in each lake. In ponds, Chaoborus were collected
during the day using a 2-L plastic pitcher dipped just below the pond surface and then
poured through a 153-micron sieve. The dipping was repeated two times per sample and
each sample was taken from equally spaced locations along a qualitative transect that
started ~0.5 m in from the pond edge to the center of each pond for a total of three 8-L

samples. C haoborus were preserved in 95% ethanol, later identiﬁed to species

 

 

8a... 9?... =88 SIN .. 882 3.8.8.3.: 2.2.3.8.? n93 925.
8.8 as .288 2?... .88 N8“ .228 a?

82 6?... a: 9.2: 2 88? 3.2.2.33 2.8.95? ".83 «5:.
.883... .38 88° .88 88 A8322

32. 82.8 a: an...” n o8? 3.8.3.3....» 2.88.??? x25
.88.: 62.8 82. 2.8:? .308 no?

82. 3?... :36 883 8: 88." 3:38.: 2.2.8.89? 8355
.88 :2. .888 ES .388... 8.3.3.:

88» 3?... 83. 888 o 882 31.930: 2.3.2.8.? 3.88253
.888 8;. .888 2.3 .288 S... $8.82....—

288 :2... G8? 388 no a? 3:88.80: 2.838.? 85882
.3826 .888 88° .38 $.u 6828 2.:

2: can... :80 883 n 83 3.8.8.8.: 2.2.8.8.? .26
6.883;. .?~.8 88° 58.8 R? .223 ~8n~

G88 38... :8: $88 82 38 3:83.53 2.8?59? :53
.8838 .888 88.8 .28 no... .868 a?

2.89 58.. a: 288 he 28 3:88.80: 2.2.2.8.? 355
.8828 .888 88° .8888 8888.8

32. 3?... G: :8: N 8? 3.8.8.803 2.3.2.3.? 255
.3388 .288 88c .38 a: 828?..."

83 $88 882 3.8 8° 2:. 3:8?393 2.82.80? 83
4.8886 $8888.. .882... 888?:

mm 253$ 33.8 88 285.!” can. 2:. aims: 8.55 3588.2

336.630 .8908 85.8. 52 8.3m
Essie

 

1012— E 05 95 «on: v:- inon >39 95.: @2338 83-? ioﬁoiﬁumuo “33558.— no noun! v:- a-uz AN 93:.

10

915.com 25 5 38>nxﬁ_&88233 0: ob? 9.05. o.

§38~ £0.82 N :Z . 8832.2... gnaw-c882

nuz.§§8§&82§uu2 .5333 dun-H.392: 3.583 NS
”aSUnSoSpu Bmuonm ans-n 30:58 EB 9253 .3 352520.. 983 in “53an 833A

 

:2 .5 .3 .2 63-2.8 8.2 8828.8 8.». 9886888 on .2 m sea-H 85:.
:2 .5 .3 .2 28.??8 282 982-88 38 98.8.8.3 8.8 N .33 82-2.
:2 .02 .3 : 8.2-8.8 8.2 3.8-3.8 8.? 388-388 8.8 :25

.5 «2 9888.8 8.8 so 8-2.8 8." 682-8. 2 8 R2 8553

.5 .3 :2 682-2 .8 9% 2.2.8.3 88 2.2.8.3 8.: 898953

.. : 82.2-8.8 Ra 682.28 8.2 6882-883 8.: 8:882

z : “2.8-2.8 8.8 8888.8 8.,“ 5.8.8.6 :2 2:6

: : 5.2-2.8 a? 288-228 8.8 5.8-2.88 8.8 28:

: : 888.88 2.2 9832-828 ?2 28.88.8288 8.28 85:

: : 288-88 8." 382.88 98 9282.88 ~82. 2x5

: : $88.88 2.2 62.2-8.8 2.? 2882-888 8.2 83

 

 

 

88.5853 832:2 383 26 a1 8v 385 36 .58. 335 253838.2- 08.5.8.2

 

82580 2.8 as:

11

(Cook 1956, Saether 1970) and measured under 40X magniﬁcation using a digitizer tablet
and software (SigmaScan Pro Version 4.01; SPSS Inc., 1987).

Pond water (1 L) was collected at each sample point along the transect using 500
ml polyethylene bottles and lake water (1 L) was collected with an integrated tube
sampler (5 cm diameter plastic tubing) that sampled the entire water column at the
deepest point of each lake. These water samples were immediately placed on ice in a
cooler, for later analysis of chlorophyll a (chl a) and total phosphorus (TP). The water
collected for the chl a analysis was divided in two and half was ﬁltered onto Gelman A/E
glass ﬁber ﬁlters (Gelman Sciences, Ann Arbor, MI USA) to measure total chl a (an
approximate for algal biomass). The other half was ﬁrst ﬁltered through a 35 um mesh
and then ﬁltered onto A/E ﬁlters to measure the “edible” size fraction. Algal size is
known to be an important feeding constraint and the <35 um fraction represents a
commonly used measure of “edible algae” (Mittelbach et al. 2004) based on known size
preferences of many ﬁlter feeding zooplankton (Sterner 1989). Ch] a was then extracted
from the ﬁlters overnight in cold 95% ethanol and measured using narrow band
ﬂuorometry (Welschmeyer 1994). Pond and lake water samples for TP measurement
were frozen and later analyzed using spectrophotometry and standard methods (Wetzel
and Likens 1991). Water temperature, dissolved oxygen, pH, and conductivity were also
measured in each pond and lake every time that Chaoborus were sampled, using a
Hydrolab® multi-probe (MiniSonde 4a). For the ponds these measures were taken at two
points, one about 0.5 m away from shore and the other near the center of each pond and
these values were averaged. For the lakes, abiotic measurements were sampled at the lake
surface and then at every meter until I reached the lake bottom and these values were
averaged for the following analyses.

Additionally, I explored the abundance of Chaoborus species in a long-term data
set (1989-2004) from Wintergreen Lake (Mittelbach et al 1995, 2006). Wintergreen Lake

experienced a gradual decrease and then increase in planktivore density over a 16-year

12

period due to the elimination of the two dominant ﬁsh species (largemouth bass:
Micropterus salmoides and bluegill sunﬁsh: Lepomis macrochirus), followed by the
subsequent intentional reintroduction of both species. For detailed methods on
zooplankton collection and estimation of planktivore densities see Mittelbach et al.
(2006).

Statistical analyses

All data presented on C haoborus species abundance includes combined counts of
III and IV instars. I used multivariate analyses to determine the important factors
inﬂuencing Chaoborus species distribution and abundance using the ﬁeld survey data.
Similarities among surveyed sites based on environmental factors (e. g. water
temperature, lake/pond area and depth, pH, conductivity, dissolved oxygen, total and <35
um chl a) and changes in Chaoborus community composition were visualized using non-
metric multi-dimensional scaling (NMDS) based on Bray-Curtis distances of the seasonal
mean Chaoborus species abundances. All multivariate procedures were performed using
PC-ORD (v4.25, McCune and Mefford 1999) and followed methods outlined by McCune
and Grace (2002).

For the long-term data set, linear regression was used to examine the relationship
between C haoborus species abundance and planktivore density. Yearly mean Chaoborus
species abundance and planktivore density were loglo (x+0.00039 (one half the lowest
density)) and logic (x) transformed, respectively to meet assumptions of the analyses.
Regression analyses were performed using Systat 11.0 (SPSS Inc., 2004).

Results

C haoborus species distributions showed remarkably clear separation when
aquatic systems were divided into ﬁsh and ﬁshless environments and arranged along a
gradient of lake depth (Fig. 2.1). Chaoborus americanus was only found in ponds without
ﬁsh; C. albatus was only found in shallow lakes and ponds with ﬁsh (2-6 m depth); C.

punctipennis was found in all lakes and ponds with ﬁsh; and C. ﬂavicans was found in

13

both ﬁsh and ﬁshless lakes and ponds (Fig. 2.1). A summary of the means and ranges for
several environmental variables measured in each lake and pond is given in Table 2.1.
With regard to relative abundance, C. americanus was by far the most abundant species
in all ﬁshless ponds and showed relatively little change in density across the sampling
period (Fig. 2.3). Chaoborus punctipenm's dominated or co-dominated the Chaoborus
assemblage of all lakes with planktivorous ﬁsh and its density remained steady or
increased through the growing season (Fig. 2.2). Chaoborus albatus and C. ﬂavicans
were more intermittent in their distributions than C. americanus or C. punctipennis and
were generally rarer than the other two species (Figs. 2.2 and 2.3). Importantly, there was
no evidence for any seasonal partitioning of these environments by the four Chaoborus
species (Figs. 2.2 and 2.3).

An NMDS ordination of the abundance of the four Chaoborus species converged
on two axes, with the ﬁnal solution representing a cumulative 94% of the variation in the
dataset (the two axes accounted for 79% and 15% of the variation, respectively). In the
NMDS plot (Fig. 2.4) there is a clear separation of two groups along axis 1, which is
strongly correlated with conductivity and pH (Pearson’s r2=0.58 and 0.60). The
separation of sites along axis 2 is strongly related to pH, depth and area (Pearson’s
r2=0.51, 0.54, and 0.87, respectively). Thus, C. americanus abundance is associated with
ponds that are shallow, have small areas, low conductivity and are somewhat acidic (Fig.
2.4, lower right). Chaoborus punctipennis and C. ﬂavicans abundances on the other hand,
are positively related to deep lakes with large areas, high conductivity and neutral to
slightly basic pH (Fig. 2.4, upper left). Chaoborus albatus abundance was linked to

shallow lakes with small surface areas (Fig. 2.4, mid-leﬁ).

14

 

 

 

 

“ v C. americanus

_ .. . . .. . .. . 0 C. punctlpennls
A C. albatus

* D C. flavicans

1 I: Cl [313 C] DD 13 C] E]

- ALLA

Chaoborus presence

 

 

I l I I I I I T I I I I
15 1411.510 6 4 3 3.53 2 1.3 0.7 0.5

Fish Fishless
Depth (m)

 

Figure 2.1. Presence of four Chaoborus species in 17 lakes and ponds that vary in

presence/absence of ﬁsh and depth.

15

101

 

 

 

 

 

 

 

 

 

 

 

 

 

Duck Lake
1 .
——v——— C. americanus
O C. punctipennis
0'1 ‘ A f —-+-- C. albatus
/ — {J -— C. ﬂavicans
00 { \x o
1 ///
.' ‘. .
0.001 1
0.0001 . . . . . , e ea
10 ‘ 1 .
,_ Three Lakes 3 Wmtergreen Lake
.2
_ . O
1... 1 . .
(D
D.
5 .
n 0.1. 0 _.,. g -o
E & i ' /‘ / \ //‘-~‘_‘
3 ‘‘‘‘‘ \ /‘\‘ / \ /
C .— TY ‘ / \ //
E D o “ /
O / CL
C
0
0.0001 I I I I I I I I *1 '7
1° ‘ Three Lakes 2 “ Lawrence Lake
1 ~ CL ~
0 » _ / "D
I} —— — 5/ U.
01 " -1 .
. \ _. o‘ ,
\. C
0.01 1 b -
D
/ U
0.001 ~ I CI”
00001 I I I I I I I I 1 I I I I I I I I I
100 120 140 160 180 200 220 240 260 280100 120 140 160 180 200 220 240 260 28'
Julian Day

Figure 2.2. Seasonal dynamics of Chaoborus densities in ﬁve southwestern Michigan

lakes (mean densities based on three samples per date).

16

 

 

 

 

 

 

 

 

100 ~ .
CIrcle Pond l Lux 18
10 'l 1
1 « l V//V\V/V\v
0.1 1 . 1
———v— C. amencanus
1. - - C. punctipennis .
0.01 . . ——+—- C. albatus
C] — {3 — C. ﬂavicans
0.001 . . . . . 4 . . . . . -
'b 100 ~
:2 Horseshoe pond l LUX 1°
5
Q 1° ‘ W 1
h
(D
.Q 1 . .
§
/CJ
c
a) 01 l U D
E
0
Q 0.01 1 ,
O
(U
-:
O 0.001
100 - .
Lux 9 LUX 16
10 j ‘ W
1 + W ‘
0 1 [I]
0.01 4 1
0.001

 

 

 

140 160 180 200 220 240 260 140 160 180 200 220 240 260

Julian Day

Figure 2.3. Seasonal dynamics of C haoborus densities in six southwestern Michigan

ﬁshless ponds (mean densities based on three samples per date).

17

 

A Three Lakes 2

Wintergreen

area
A?! X C. ﬂavicans
C. punctipennis

L".- ,1 », depth
Three Lakes 3
Lawrence \
A X C. albatus
ed;ble :.:i"2a0I-ophyll a
T D C. americanus
I Horseshoe‘ LUXI6
LuxlO
Circle
Duck tenmeratu'e ng

A ' A Lux18

 

 

 

Figure 2.4. Association of 5 lakes and 6 ponds based on NMDS ordination of Chaoborus
species abundances. The angle and length of the joint plot lines indicate the direction and
strength of the relationship of environmental variables with the ordination scores.

Asterisks represent the centroids for each Chaoborus species.

18

The ﬁeld patterns indicate that C. punctipennis tends to dominate in all systems
with planktivorous ﬁsh and that C. albatus might beneﬁt under conditions of increased
planktivory while C. ﬂavicans should be neutral to changes in planktivore density.
Looking at a l6-year record from Wintergreen Lake, where planktivore densities have
varied across two orders of magnitude due to the elimination and reintroduction of ﬁsh
species (Mittelbach et al. 2006), shows responses consistent with these expectations.

C haoborus punctipennis was consistently the dominant Chaoborus species in the lake
and planktivore density had no signiﬁcant effect on its density (Fig. 2.5; r2=0.08, p=0.36).
Chaoborus albatus responded positively to increased planktivore density (Fig. 2.5;
r2=0.31, p=0.046) and planktivore density had no signiﬁcant effect on C. ﬂavicans
density (Fig. 2.5; r2=0.08, p=0.36).

Discussion

Fish presence or absence is a major factor determining the distribution of many
aquatic taxa in lakes and ponds (Wellbom et al 1996). My survey of chaoborid
distributions clearly documents the inﬂuence of ﬁsh on the distribution and abundance in
this assemblage. Chaoborus americanus was present only in ﬁshless systems and in these
systems it was always the numerical dominant. In fact, it was rare to ﬁnd any other
Chaoborus species co-occurring with C. americanus. C. punctipennis, conversely, occurs
only with ﬁsh.

Previous observational studies have observed similar patterns in Chaoborus
distribution where C. americanus was found only in ﬁshless water bodies and C.
punctipenm's was positively correlated to the presence and sometimes intensity of ﬁsh
planktivory (von Ende 1979, Ramcharan et al. 2001, Wissel et al. 2003). Such patterns
may be the result of the direct effect of ﬁsh predation or could be a consequence of the
indirect effects of ﬁsh on the shared zooplankton prey. Planktivorous ﬁsh feed selectively
on the largest prey including Chaoborus (Garcia, Chap.4). Thus, the absence of C.

americanus in ﬁshless systems could be due to size selective predation by ﬁsh since

19

 

10

 

 

 

 

 

 

 

 

A 1 "
d
4:1:
v
3‘ 0.1 -1
'17)
c
(D
‘O 0.01 ‘
S
O
.8 0.001 4
m o o D
5 O C. punctlpennls

0-0001 - A C. albatus [3

C] C. ﬂavicans
0.00001 1 1 1
100 1000 10000 100000 1000000

Planktivore density (# per lake)

Figure 2.5. C haoborus density in Wintergreen Lake as a function of variation in
planktivore density over 16 yrs. Lines are the ﬁtted linear regressions for each species
(the solid line is signiﬁcant, p<0.05). Planktivore densities determined by mark/recapture

(see Mittelbach et al. 2006 for details).

20

C. americanus is the largest of the four chaoborid species. However, planktivory may
also indirectly affect C haoborus competitive interactions through changes in their
zooplankton resources. There is abundant evidence that fish shift the size structure of the
zooplankton community towards smaller-bodied prey (Brooks and Dodson 1965, Gliwicz
and Pijanowska 1989). Because Chaoborus are gape-limited predators, this change in
prey size structure may be advantageous for the smaller-bodied Chaoborus species (C.
punctipenm's and C. albatus) and thereby enhance recruitment of these species. However,
there was no change in C. punctipennis dominance and abundance in Wintergreen Lake
despite >2-fold changes in mean cladoceran body length and species composition related
to the gradual decrease and then increase in planktivore density through time (Mittelbach
et al. 2006). In addition, the Chaoborus species differ in DVM behavior, C. punctipennis,
C. ﬂavicans and C. albatus exhibit DVM behavior, whereas C. americanus does not
(Berendonk et al. 2003). Thus, C. americanus is especially vulnerable to planktivorous
ﬁsh (Garcia, Chap. 4) and a strong case can be made for the importance of the direct
effect of ﬁsh predation in excluding the C. americanus from environments with ﬁsh.

Chaoborusﬂavicans does not seem to discriminate between habitats with and
without ﬁsh (Fig. 2.1) and this is supported by the literature as well (Pope et al 1973,
Gonzalez and Tessier 1997, Wissel et al. 2003). Instead, Wissel et al. (2003) show that
the abundance of C. ﬂavicans was higher in lakes that were relatively small and that had
elevated levels of dissolved organic carbon (DOC). My NMDS ordination shows a
somewhat contradictory pattern, in that C. ﬂavicans abundance was positively related to
lake area, as well as conductivity and higher pH (Fig. 2.4). Lake area may increase or
decrease the intensity of planktivory. Therefore, direct manipulations of planktivore
density are needed to sort out their effects on C. ﬂavicans abundance.

Chaoborus albatus showed a very clear pattern of only being found in shallow
lakes with ﬁsh. This result indirectly implies that the distribution of this species may be

dependent not just on the presence of ﬁsh predators but on the level of ﬁsh predation

21

pressure. That is, environmental factors such as area and depth are considered to be
related to the intensity of ﬁsh planktivory; shallow lakes tend to support higher levels of
ﬁsh planktivory, due to the lack of a hypolimnentic refuge and because piscivores are
oftentimes absent (Tessier and Woodruff 2002). In addition, I observed that within
Wintergreen Lake, where lake area, depth and other environmental variables remained
relatively consistent but planktivory changed dramatically over a l6-year period, the
abundance of C. albatus increased signiﬁcantly with increasing ﬁsh planktivory (Fig.
2.5).

From the ﬁeld pattern and ﬁeld survey we see that C. americanus never co-
occurred with C. punctipennis or C. albatus, but was often found together with C.
ﬂavicans (Fig. 2.1). This indicates that competition and possibly predation by larger
Chaoborus species could be another factor contributing to the pattern in the Chaoborus
species assemblage across the landscape. von Ende (1979) observed that late instars of C.
americanus will prey on the smaller C. punctipennis and proposes this as a possible
mechanism excluding C. punctipennis from systems where C. americanus is present. In
addition, the low incidence of coexistence between species of C haoborus found in prior
studies has been attributed to competitive exclusion since there is substantial overlap in
resource use between species of C haoborus. Thus, when coexistence is observed it is
often ascribed to differences in phenology (Roth 1968, Carter and Kwik 1976, von Ende
1982, Sardella and Carter 1983). However, in my ﬁeld survey, 1 found little evidence of
seasonal partitioning of the environment by the different Chaoborus species. Instead, the
four species showed similar patterns of abundance across the season, at least in the III
and IV instar stages measured in this study. Thus, there was no clear evidence for
coexistence via temporal resource partitioning. Other possibilities such as differences in
depth distributions, prey selection, etc. remain to be explored.

The results of this observational study provide signiﬁcant insight into the

importance of ﬁsh predation in patterning the Chaoborus assemblage, although the exact

22

mechanism by which sorting occurs remains unknown. Chaoborus species may exhibit
oviposition habitat selection and thus select or avoid environments with ﬁsh.
Alternatively, dispersal and oviposition may be random, and distribution patterns may be
generated by differential mortality and species sorting. This can be tested by direct

experimental manipulation of ﬁsh densities across a landscape of ponds and lakes.

23

CHAPTER 3

OVIPOSITION HABITAT SELECTION AND LARVAL PERFORMANCE

Abstract
The distribution of organisms across the landscape may be a function of random
dispersal followed by non-random, site-speciﬁc mortality (species sorting). Alternatively,
species may show directed dispersal and habitat selection. In southwestern Michigan,
four species of Chaoborus (Diptera: Chaoboridae) show pronounced differences in their
distribution in relation to variation in ﬁsh density. Here, I examine whether Chaoborus
species that differ in their vulnerabilities to ﬁsh predators, discriminate between potential
oviposition sites based on the presence of ﬁsh and ﬁsh cues. As their vulnerabilities
would suggest, C. americanus, but not C. punctipennis or C. ﬂavicans, showed evidence
of oviposition habitat selection for ﬁsh-free sites in a ﬁeld mesocosm experiment.
However, in a second experiment, C. americanus did not avoid ovipositing in habitats
with ﬁsh cues. This conﬂicting result may be due to possible strong site ﬁdelity in this
species. Oviposition habitat selection, together with direct consumption by ﬁsh and
possibly larger Chaoborus species, as well as competitive interactions between species of
C haoborus appear to be important mechanisms inﬂuencing the distribution of C haoborus
species across the landscape.
Introduction
‘ It is widely appreciated that the presence of predators can directly and/or

indirectly affect the abundance and distribution of prey species, particularly in aquatic
communities (Kerfoot and Sih 1987, Zaret 1980, Wellbom et al. 1996). In response to
predators, prey have evolved many adaptations to reduce their mortality risk such as diel
vertical migration behavior, morphological changes, and chemical defenses (Kerfoot and
Sih 1987). While these adaptations are important factors structuring communities, the
question of whether they lead to the distributional patterns of prey is unclear (Binckley

and Resetarits 2003).

24

A growing body of evidence suggests that anurans and some insects can use
chemical signals produced by predators to assess the risk of predation in a habitat
(Blaustein 1999, Resetarits 2001, Abjornsson 2002). Oviposition habitat selection (OHS)
in response to ﬁsh predators can strongly affect individual ﬁtness and this may be an
important mechanism generating the presence/absence patterns of prey abundance
(Blaustein 1999, Resetarits 2001). Natural selection should favor females that avoid
ovipositing in habitats where risk of predation for their offspring is high (Blaustein et al.
2004). Differences in vulnerability to ﬁsh predation, due to prey adaptations or prey size,
may select for OHS behavior for some prey species and not for others.

In this study, I examined whether OHS in response to ﬁsh predators accounts for
the distribution of Chaoborus (Diptera: Chaoboridae) species across a landscape of ﬁsh
and ﬁshless ponds. Most studies examining the processes inﬂuencing the distributional
pattern of Chaoborus species across habitats have focused on the direct effects of ﬁsh
predators (von Ende 1979, Wissel et al. 2003); empirical tests investigating indirect (non—
lethal) effects are rare.

Larvae of the phantom midge Chaoborus are common inhabitants of the plankton
of North American lakes and small ponds. Of the 12 described Chaoborus species in
North America (Seether 1970), four are found in southwestern Michigan: C. americanus,
C. punctipennis, C. ﬂavicans, and C. albatus. All Chaoborus species develop through
four aquatic larval instars before becoming pupa, followed by a short lived, non-feeding
ﬂying adult stage where females lay one egg raﬁ per lifetime (Moore 1986, Borkent
1979). Egg rafts may contain over a hundred eggs (Saether 1997). Chaoborus larvae are
gape-limited, ambush predators of small to medium sized zooplankton (e.g., Moore et a1.
1994, Swift and Fedorenko 1975, Pastorok 1981) and are vulnerable to predation by
planktivorous ﬁsh (Garcia, Chap.4). These four species vary widely in morphology from
the relatively large, opaque C. americanus (4th instar length 10-13 mm) to the similar-

sized, but transparent C. ﬂavicans (4‘h instar length 9-12.65 mm) to the small, transparent

25

C. punctipennis (4‘h instar length 7.5-9.5 mm) and C. albatus (4‘h instar length 7-9.4 mm)
(Cook 1956, Sazther 1970). C. punctipennis, C. albatus, and C. ﬂavicans are most often
found in lakes with ﬁsh and exhibit diel vertical migration (DVM) as an adaptive
response to the presence of ﬁsh in these habitats (Dawidowicz et al 1990, Tjossem 1990).
Chaoborus americanus is generally observed in ﬁshless ponds and does not exhibit DVM
behavior (Berendonk et al. 2003). Also, C. americanus rarely coexists with other species
of Chaoborus (Pope et al. 1973, von Ende 1979, Wissel et al. 2003). The lack of co-
occurrence of C. americanus and other species of Chaoborus may be the result of
interspeciﬁc competition or intraguild predation, or may reﬂect differences in oviposition
habitat selection. Chaoborus americanus is the most vulnerable of the four species to ﬁsh
predation (Garcia, Chap. 4). Therefore I hypothesized that: (la) C. americanus will
exhibit oviposition habitat selection because risk of predation is high for their progeny,
but (lb) C. ﬂavicans, C. punctipennis, and C. albatus will not exhibit OHS because they
have evolved a behavioral response to the presence of ﬁsh; and (2) the presence of C.
americanus in a habitat will exclude the other three species.

I tested these hypotheses in two similar experiments. The ﬁrst was an outdoor
mesocosm experiment allowing ovipositing Chaoborus to select between large tanks with
free-swimming ﬁsh, caged ﬁsh (ﬁsh cue) and controls without ﬁsh. The second was an
enclosure (bag) experiment conducted in a ﬁshless reservoir, where C haoborus could
choose between bags with a single caged ﬁsh (ﬁsh cue), two caged ﬁsh (double ﬁsh cue)
and controls without ﬁsh.

Methods

The two experiments were conducted at the W. K. Kellogg Biological Station
(KBS), Experimental Pond Facility (Hickory Comers, MI). Experiment 1 (mesocosm
experiment) was performed from June-September 2004 in 24, aquatic mesocosms (300 L,
1 m diameter, cattle tanks). On 28 May 2004, the mesocosms were acid-washed, ﬁlled

with low nutrient well water, and equipped with a free ﬂoating ﬁsh cage (2 mm mesh, 25

26

cm diameter). The experimental design consisted of four treatments: 1) control (without
ﬁsh), 2) high productivity with no ﬁsh, 3) ﬁsh cue (one caged ﬁsh), and 4) ﬁsh (one free-
swimming ﬁsh), each replicated six times. Phosphorus (as potassium phosphate) and
nitrogen (as ammonium nitrate) were added to the water column of all tanks to bring
nutrient concentrations to levels commonly observed in local lakes and ponds (Darcy-
Hall 2006). Target nutrient supply concentrations were 25 (for control, ﬁsh and ﬁsh cue
tanks) and 100 (for high productivity tanks) ug/L total phosphorus (TP), with nitrogen
added in a 50:1 molar ratio. Nutrients were added the day the tanks were ﬁlled and every
10 days after that throughout the experiment to maintain approximate target nutrient
concentrations.

A diverse algal inoculum collected from eight local ponds was introduced into
each tank two days after they were ﬁlled. Ten snails (Physa sp.) were also added to each
tank to regulate periphyton growth on the bottom and sides of the tank. A week later, a
diverse zooplankton inoculum collected from the same eight local ponds was introduced
into each tank after samples were ﬁltered to remove all potential predators of Chaoborus.
Zooplankton collected from these same ponds were added every ten days throughout the
experiment to maintain prey populations. The tanks were placed in 2 rows separated by
1.5 m, with treatments and replicates randomly assigned. Bluegill sunﬁsh (Lepomis
macrochirus) (32-35 mm SL) collected from one of the experimental ponds were added
to the appropriate treatments on 22 June 2004. This bluegill stocking density (~1.3
bluegills/mz) is within the natural range of bluegill densities found in this region
(Mittelbach 1988). Tanks were covered with ﬁberglass screen lids until ﬁsh were added
and thereafter were left uncovered, open to colonization. The presence and number of
Chaoborus egg rafts were checked daily throughout the experiment, although egg rafts
could not be identiﬁed to species. C haoborus larvae were sampled weekly with a dip net
(0.25 m diameter, 250-micron mesh), by two sweeps around the circumference of the

tank, just under the surface and again at the bottom. A water column sample was also

27

taken using a round sieve (50 cm diameter, 500-micron). Chaoborus larvae were counted,
identiﬁed to species and promptly returned to their original tank. All sampling equipment
was washed with well water and a power nozzle between tanks to reduce risk of
contamination. On 15 September 2004, ﬁsh were removed and all tanks were drained
completely and ﬁltered through a sieve (500-micron). Chaoborus larvae were preserved
in 75% ethanol, later enumerated under 40X magniﬁcation, and identiﬁed to species.
Water temperature, dissolved oxygen, pH, and conductivity were measured in each tank
every time that Chaoborus were sampled, using a Hydrolab® multi-probe (MiniSonde
4a).

Experiment 2 (bag experiment) was performed from July-September 2005, in a
ﬁshless reservoir using 1200-liter polyethylene “bag” enclosures (1m diameter, 1.5m
deep), sealed at their bottoms and suspended in the water column from ﬂoating frames.
Bags were equipped with the free ﬂoating ﬁsh cages used in the mesocosm experiment,
and covered with ﬁberglass screens until ﬁsh introduction. To explore effects of ﬁsh cue,
three treatments were employed: 1) control (without ﬁsh), 2) ﬁsh cue (one caged ﬁsh)
and 3) double ﬁsh cue (two caged ﬁsh). Each treatment was replicated six times and
randomly assigned for a total of 18 enclosures. Bags were ﬁlled by pumping water from
the reservoir through a 153-micron zooplankton net to remove large zooplankton and
invertebrate predators. A diverse zooplankton inoculum collected from six local ponds
was introduced into each bag after samples were ﬁltered to remove all potential predators
of Chaoborus. A month later, 1 August 2005, bluegill sunﬁsh (54-70 mm SL) from one
of the experimental ponds were added to the appropriate treatments. The bluegill stocking
densities (l and 2 bluegills/mz) are within the natural range of bluegill densities found in
this region (Mittelbach 1988). Bags were then left uncovered, open to colonization by
ovipositing Chaoborus. On 26 September 2005, the bags were sampled by taking ﬁve
vertical tows through the entire water column of each bag with a 30 cm diameter, 500-

micron plankton net. Chaoborus were preserved in 75% ethanol, measured and identiﬁed

28

to species (Cook 1956, Saether 1970) under 40X magniﬁcation using a digitizer tablet and
software (SigmaScan Pro Version 4.01; SPSS Inc., 1987). Additionally, water
temperature, dissolved oxygen, pH, and conductivity were measured using a Hydrolab®
multi-probe (MiniSonde 4a).
Statistical Analyses

Examination of treatment effects for the mesocosm experiment and the bag
experiment was performed using repeated measures ANOVA (rmANOVA) and one—way
ANOVA, respectively. For the mesocosm experiment, Chaoborus species abundances
per replicate were averaged across two consecutive sampling dates. In addition, values
for the experiments were loglo (x+(one half the lowest observed density); mesocosm =
.0017 and bag = 0.0016)) transformed to meet assumptions of the analyses. Statistics
were performed using Systat 11.0 (SPSS Inc., 2004).
Results
Mesocosm experiment

The ﬁrst Chaoborus egg raft appeared (in a ﬁsh cue tank) one week after the start
of the experiment and egg rafts were found at least once in all treatments throughout the
experiment. C haoborus americanus, C. punctipennis and C. albatus larvae ﬁrst appeared
in the control and high productivity treatments two weeks into the experiment. The
rmANOVA revealed a signiﬁcant treatment effect only for C. americanus (p=0.007,
F3,20=5.32, between subjects effect), where C. americanus abundance was higher in the
ﬁshless treatments (Fig. 3.1, control and high productivity). Because there were no
signiﬁcant effects of time or time by treatment interactions, the experimental data are
shown as Chaoborus density averaged across all sampling dates within each of the four
treatments (Fig. 3.1). Closely reexamining the ﬁshless treatments showed that if C.
americanus was present in the tank then the other two Chaoborus species were not (Fig.

3.2a) and if C. ﬂavicans was present in a tank then C. punctipennis was not (Fig. 3.2b).

29

0.1

 

 

 

 

A C. albatus
V C. americanus :7
3 0011 E] C. flavicans
E O C. punctipennis
.39
co _
5 0.001 %
‘O
(I)
E 0.0001 1 El
0
.Q
o
E
0 0.00001 -
0.000001 . . I 1

Fish Cue Fish Control High Productivity

Figure 3.1. Effects of ﬁsh cue, ﬁsh, and high productivity on mean Chaoborus density

(number per liter 3: 1 SE) averaged across all sampling dates.

30

0.06

 

 

 

 

 

 

 

 

 

 

a) 0 Control
0-05 ‘ 9 0 High Nutrient
O
004 ~
._I O
R
L 0.03 1
2
w 0
9 0.021
0 O
O O
0.01 « O
8
0.00 . 000 O O
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
C. americanus #IL
0.06
b)
0.05 «
O
< 0.04 1
=14: O
‘3 0.03 -
(U
.9
a 0.021
“F O
0 0.01 1
O
O
0.00 — O O O O O

 

 

0.00 0.01 0.02 0.03 0.04 0.05 0.06
C. punctipennis #lL

Figure 3.2. C haoborus density (number per liter) in the ﬁshless treatments of the
mesocosm experiment. Each point is a sampling date for an individual tank. (C. other

includes C. punctipennis and C. ﬂavicans).

31

Lastly, C haoborus failed to pupate in the treatments with ﬁsh, but pupa of all three
Chaoborus species were found in the ﬁshless treatments.
Bag experiment

This experiment differed from the mesocosm experiment in three main ways: 1)
the environmental matrix surrounding the treatments, 2) scale (bags were deeper and
wider than the tanks), and 3) the focus on the effect of the intensity of ﬁsh cue on
Chaoborus oviposition decisions. The bags were placed in a reservoir where C.
americanus was present, whereas in the mesocosm experiment the tanks were in a
mowed ﬁeld at the experimental pond facility. The ﬁrst egg rafts appeared two days after
the start of the experiment in at least one bag of each treatment. Chaoborus americanus
was observed in all treatments, C. ﬂavicans was found in the double ﬁsh cue and control
bags and C. punctipennis was observed in one of the double ﬁsh cue bags (Fig. 3.3).
There was only a signiﬁcant effect of the intensity of ﬁsh cue on C. americanus, where
C. americanus abundance was highest in the double ﬁsh cue and control treatments (Fig.
3.3; p=0.04, F2,.5= 4.23, one-way ANOVA).
Discussion

Several recent studies have observed oviposition habitat selection (OHS) in
response to risk of predation (Blaustein 1999, Blaustein et al. 2004, Rieger et al 2004).
Blaustein (1999) suggested that the evolution of this behavior is most likely under the
following conditions: 1) immature stages are highly vulnerable to the predator; 2)
predator density is highly variable among patches; 3) prey have few lifetime reproductive
events; and 4) predator distribution among patches is largely ﬁxed from the time the
female oviposits until offspring leave the patch. Chaoborus americanus meets all of the
conditions that should favor OHS and in the mesocosm study I found evidence that C.

americanus exhibited OHS. However, when the ﬁsh-free and ﬁsh cue treatments were

32

 

0.1

 

 

 

A
€ 0.01 . §
‘5’ i
5 I 9
c 0.001
m
'U
3
L 0.0001 -
0
Q
o .
g V C. americanus
0 0.00001 . Cl C. ﬂavicans
O C. punctipennis
0.000001 . . .
Control Fish Cue 2X Fish Cue

Figure 3.3. Effects of ﬁsh cue on mean density (number per liter :t 1 SE) of three

C haoborus species on the ﬁnal day of the bag experiment.

33

located in bags suspended in a ﬁsh-free reservoir that supported a population of C.
americanus, C. americanus did not discriminate between ﬁsh cue and ﬁshless treatments.
Chaoborus punctipennis and C. ﬂavicans do not meet the condition of high larval
vulnerability to the predator (Dawidowicz et al 1990, Tjossem 1990) and did not avoid
ﬁsh or ﬁsh cue treatments in either experiment.

Berendonk (1999) found a similar pattern where the Chaoborus species most
vulnerable to ﬁsh predation avoided ovipositing in containers with ﬁsh kairomones (C.
crystallinus and C. obscuripes, both Old World pond-dwelling species) and the less
vulnerable C. ﬂavicans was not observed to discriminate between ﬁsh kairomone and
control containers. These results support the suggestion that the avoidance of ovipositing
in habitats with ﬁsh by certain C haoborus species is based on chemical and not visual
cues. Berendonk and Bonsall (2002) also found that C. crystallinus avoided ovipositing
in barrels with caged ﬁsh, but observed that this oviposition preference decreased with
distance from the source population. McPeek (1989) reported that Enallagma damselﬂies
from ﬁshless lakes could not distinguish between experimentally manipulated ﬁsh and
ﬁshless ponds and he attributed this lack of a response to the damselﬂies strong
philopatry. These studies suggest several possible explanations for the contrasting results
of the bag experiment. The ﬁrst possibility assumes that female C. americanus may have
been dispersing from distant source populations and thus may have reached a critical
point and had to oviposit regardless of the suitability of the habitat. However, the bag
experiment was conducted in a habitat with a population of C. americanus and hence
dispersal distance would have been minimal. The second and third, more likely
explanations may be that C. americanus exhibits strong site ﬁdelity, and/or that C.
americanus could not detect (or misinterpreted) the ﬁsh cue in the bags because the
matrix was “ﬁsh free”.

In the mesocosm experiment, I observed no coexistence between the Chaoborus

species. In particular, when C. americanus was present, all other species were absent.

34

This pattern is consistent with patterns observed in nature, where C. americanus is often
the only Chaoborus species in ﬁshless ponds (Pope et al. 1973, Garcia Chap.2). von Ende
(1979) hypothesized that predation by C. americanus on smaller species of Chaoborus
may be one of the mechanisms preventing coexistence among these species. If this is the
case, C. punctipennis and C. ﬂavicans may not select oviposition sites based on ﬁsh cues,
but might instead use habitat area, lake depth, or some other factor that was not
considered in this present study.

In this study, I show that oviposition habitat selection may contribute to the
distributional pattern of C. americanus occurrence across the landscape. However, from
these experiments OHS does not contribute to the distributional patterns of C.
punctipennis and C. ﬂavicans. Instead, direct consumption by ﬁsh and possibly larger
Chaoborus species as well as competitive interactions between species of Chaoborus
may be the dominant factors inﬂuencing C. punctipennis and C. ﬂavicans distribution;
this remains to be investigated both empirically and theoretically. Distribution and
abundance patterns at the community scale can result from two distinct mechanisms;
random dispersal followed by non-random, site-speciﬁc mortality (species sorting) and
oviposition habitat selection (Binckley and Resetarits 2005). For active dispersing
organisms such as C haoborus examining potential ﬁlters at the colonization stage is an

important ﬁrst step in understanding how patterns in species abundance are created.

35

CHAPTER 4

CHAOBOR US SPECIES SORTING ALONG A PREDATION GRADIENT

Abstract
Variation in the intensity of predation pressure across the well-known
environmental gradient of lentic freshwater habitats from small, ephemeral ponds to
large, permanent lakes is a key ecological interaction important in the development and
maintenance of aquatic community structure. Here, I examined Chaoborus (Diptera:
Chaoboridae) species sorting along an experimental gradient in bluegill sunﬁsh (Lepomis
macrochirus) density. In addition, I tested bluegill prey choice for the four species of
Chaoborus found in Southwestern, MI. I found that a gradient in ﬁsh density can lead to
clear species sorting of four Chaoborus species, and that this response is consistent with
distributional patterns observed in nature. Chaoborus americanus was most abundant in
the ﬁshless ponds, C. fIavicans was neutral in response to ﬁsh and C. punctipennis was
most abundant in the high ﬁsh biomass ponds. Furthermore, ﬁsh size selectivity and
differences in Chaoborus species traits (i.e. pigmentation, DVM behavior and size) were
all observed to contribute to the pattern of Chaoborus abundance and distribution.
Introduction
The pattern of species turnover along ecological gradients can reveal factors

potentially important in determining their abundance and distribution. Ecologists have
long recognized the environmental gradient of lentic freshwater habitats from small,
ephemeral ponds to large, permanent lakes as a critical axis along which aquatic
communities are organized (Wellbom et al. 1996, Stoks and McPeek 2003). Across this
gradient, species from many freshwater taxa sort out according to physical factors (e.g.
pond drying) and biotic interactions (e. g. predation; Wellborn et al. 19996, Werner and
McPeek 1994, Stoks and McPeek 2003). For example, the intensity of ﬁsh planktivory
may vary across gradients in lake size, from high levels in shallow lakes without

piscivores, to medium levels in shallow lakes with piscivores, to low levels in deep lakes

36

with piscivores and a refuge. Many species of plankton separate out across this gradient
in lake type (Tessier and Woodruff 2002). However, while the correspondence between
species distributions and environmental gradients may suggest causal relationships, the
true test of these relationships depends on experimentally manipulating the hypothesized
causal factors.

In north temperate lakes, species of the phantom midge Chaoborus (Diptera:
Chaoboridae) reveal consistent distribution patterns across the landscape (Garcia, Chap.
2), that strongly suggest the importance of ﬁsh predation and possibly other factors
known to vary across the landscape such as, interspeciﬁc interactions, water
transparency, temperature, and nutrient levels (Pope et al. 1973, von Ende 1979,
Lamontagne et al. 1994, Wissel et al. 2003). Most of the above observational studies of
Chaoborus distributions have focused on a single species or a single habitat type and few
have directly measured the abundance of the predators presumed to drive the pattern of
species turnover. A powerful way to examine the role of ﬁsh predation in determining
species distributions in this and other systems is to establish an experimental gradient in
ﬁsh density and use regression analysis examine the response of the Chaoborus species
assemblage to this gradient (Wellbom et al. 1996). This experimental design provides
invaluable information, vital to the development of simulation models, and can be used to
make predictions for new systems (e.g., Cottingham et al 2005 and references therein).

Here I examine C haoborus species sorting across an experimental gradient of
bluegill sunﬁsh (Lepomis macrochirus) density. I asked two questions: 1) How does
predator (ﬁsh) density affect Chaoborus species composition and relative abundance?
and 2) Is the pattern of prey choice by bluegill consistent with the observed pattern of
Chaoborus species sorting across the predation gradient? I tested the ﬁrst question in two
ﬁeld experiments in ponds in southwest Michigan. I tested the second question in a prey

preference experiment conducted in a large, outdoor mesocosm.

37

Methods
Study Organisms

The larvae of the phantom midge Chaoborus are distributed worldwide and are
common inhabitants of North American lakes and small ponds. All species develop
through four instars, pupate and become non-feeding ﬂying adults that lay one egg raft
per lifetime (Moore 1986, Borkent 1979). They are planktonic, gape-limited, ambush
predators of small to medium sized zooplankton (e.g., Moore et al. 1994, Swift and
Fedorenko 1975, Pastorok 1981). Of the 12 described Chaoborus species in North
America (Saether 1970), four are found in southwestern Michigan: C. americanus, C.
punctipennis, C. ﬂavicans, and C. albatus. These four species vary widely in morphology
from the relatively large, opaque C. americanus (4m instar length 10-13 mm) to the
similar-sized but transparent C. ﬂavicans (4th instar length 9-12.65 mm) to the small,
transparent C. punctipennis (4th instar length 7.5-9.5 mm) and C. albatus (4th instar length
7-9.4 mm) (Cook 1956, Seether 1970). C. americanus is found in environments without
ﬁsh and does not exhibit diel vertical migration (DVM) behavior. The other three species
occur commonly with ﬁsh and vertically migrate; they are found in the upper waters of
lakes at night and then migrate down into the sediments during the day (e. g., von Ende
1979, Tjossem 1990)

The bluegill sunﬁsh is an important planktivore in small lakes throughout the
eastern United States (e.g., Mittelbach and Osenberg 1993, Mittelbach et al. 2006,
Werner et al. 1977, 1978). They are diurnal, size selective, and prey preferentially on
large zooplankton such as C haoborus and Daphnia (Mittelbach 1981, Turner and
Mittelbach 1990).

Fish Gradient Experiments

The experiments were performed in a series of ponds (each 30m diameter and

1.6m max. depth) located at the W. K. Kellogg Biological Station (KBS) Experimental

Pond Facility in southwestern MI. These experimental ponds support an invertebrate

38

fauna characteristic of small, mesotrophic ponds in southwestern Michigan (Garcia,
Chap. 2) and all four Chaoborus species native to the region occur at the Experimental
Pond site. In the ﬁrst experiment, conducted May-Sept. 2003, I looked at the effects of
planktivore density on the potential colonization and population growth of the four
Chaoborus species. For the ﬁrst experiment, ten ponds were drained in September 2002,
all ﬁsh were removed, and the ponds remained dry through the winter. In late April/early
May (28‘h-2nd) I ﬁlled the ponds with water from the same source (a ﬁshless reservoir on
site which contained some larvae of C. americanus) and then established a gradient in
ﬁsh density by stocking adult bluegill from nearby Warner Lake on 16 May 2005 (Table
4. l A).

In a second experiment, conducted in summer 2005, I used a larger number of
ponds (15) and allowed Chaoborus and other invertebrate p0pulations to establish in the
ponds for a year before adding ﬁsh. The 15 experimental ponds were drained in
September 2003 and remained dry through the winter. In May 2004, the ponds were
ﬁlled with water from the same ﬁshless reservoir used in Experimental 1 and were left
undisturbed (and ﬁshless) for one year. I established a gradient in ﬁsh density by stocking
adult bluegill from nearby Wintergreen Lake on 24 May 2005 (Table 4.18). For both
experiments, ponds were assigned haphazardly to the gradient in ﬁsh biomass. The
standard length (SL) of each ﬁsh added was measured (Experiment 1: range 30-75mm
SL; Experimental 2: range 100-145mm SL) and total ﬁsh biomass per pond was
calculated with a length-weight regression using a subset of the ﬁsh collected that were
not used in the experiments. Stocked bluegill biomasses (Experiment 1: ~0.07-8 g/mz;
Experiment 2: ~0.2-4 g/mz) were within the range found in nearby lakes (Mittelbach and
Osenberg 1993).

Bluegill grew and reproduced in the ponds in both years. To determine ﬁnal ﬁsh
biomass in each pond at the end of the experiments, I ﬁrst captured adult and young of

the year (YOY) bluegill with a beach seine (23 x 1.8m, 3.2-mm mesh; two seines per

39

pond), then drained the pond and removed the remaining ﬁsh. All adult bluegill were
collected and measured. A random sample of 40 adults were weighed and measured to
generate length-weight regressions to calculate biomass. In addition, a majority of YOY
bluegill were collected and weighed. Final ﬁsh biomass per pond (adults plus YOY) was
used in the analyses. Similar results were found using initial bluegill biomass or the mean
of initial and ﬁnal biomass.

In Experiment 1, C haoborus and other zooplankton were sampled once before
ﬁsh addition and then every 10-14 (1 thereafter for 17 weeks. C haoborus and other
zooplankton were collected one hour after sunset by taking three Schindler trap samples
(18.5 L, 60-micron net) at the deepest point in each pond. In Experiment 2, Chaoborus
were sampled once prior to the start of the experiment and then weekly for four weeks,
and then every three weeks thereafter as changes in Chaoborus species composition
slowed. I used vertical tows with a large-mesh plankton net to collect Chaoborus in
Experiment 2 as opposed to the Schindler trap used in Experiment 1, as I wanted to
collect a greater number of Chaoborus per sampling date than I was able to collect in
Experiment 1. Chaoborus were collected one hour after sunset by taking three vertical
tows through the entire water column with a 30 cm diameter, 500-micron plankton net, at
the deepest point in each pond. Zooplankton were collected once midway through
Experiment 2 (day 50) one hour aﬁer sunset by taking three vertical tows through the
entire water column with a 30 cm diameter, 153-micron plankton net, at the deepest point
in each pond. Samples were preserved in 95% ethanol and later enumerated and
measured under 40X magniﬁcation using a digitizer tablet and software (SigmaScan Pro
Version 4.01; SPSS Inc., 1987). Chaoborus and cladocerans were identiﬁed to species,
copepods to suborder (i.e., calanoids and cyclopoids) and rotifers were also counted.
Water temperature, dissolved oxygen, pH, and conductivity were measured in each pond
every time that C haoborus were sampled, using a Hydrolab® multi-probe (MiniSonde

4a).

40

Statistical Analyses

For each ﬁsh gradient experiment, I examined the effect of ﬁsh biomass on
Chaoborus species abundance for the period where Chaoborus densities had stabilized
(Experiment 1: days 56-123; Experiment 2: days 50-112), using repeated measures
ANOVA (rmANOVA), grouping ﬁsh biomass into three categories; zero ﬁsh biomass,
medium biomass (Experiment 112300-3200 g; Experiment 2:500-1250 g) and high
biomass (Experiment 1:5700-12,050 g; Experiment 2:3250-5850 g). For each pond,
Chaoborus species abundances were averaged across three Schindler trap samples per
date (Experiment 1) or across three vertical tows per date (Experiment 2) and then loglo
(x+(one half the lowest observed density); Experiment 1 = 0.00393 and Experiment 2 =
000442)) transformed to meet assumptions of the analyses. Linear regression was used to
examine: 1) the relationship between Chaoborus species density (Experiment 1: mean of
sampling days 56-123; Experiment 2: mean of sampling days 50-112) and ﬁnal ﬁsh
biomass, and for Experiment 2 only, 2) the relationship between ﬁnal ﬁsh biomass and
number of days post ﬁsh addition that C. americanus was present in the water column of
each pond.
Prey Preference Trials: Non-refuge and refuge

Eight feeding selectivity experiments using 1000-L cattle tanks were performed in
2005 to determine bluegill preference for three of the four Chaoborus species common to
southwestern Michigan and occurring in the ﬁsh gradient experiments (C. albatus was
too rare to use in the feeding trials). Cattle tanks were acid—washed and ﬁlled with well
water prior to the initiation of the experiments. C. punctipennis and C. ﬂavicans were
collected from two nearby lakes where they are common (Little Mill and Bristol) and C.
americanus was collected from two nearby ponds (Lux l6 and Pond A). Field collections
of Chaoborus and other zooplankton were left overnight in 16 liter buckets with the lids
on to eliminate most of the non-Chaoborus zooplankton and then ﬁltered through a 1 mm

sieve the next day. All three species of Chaoborus were stocked into a 1000-L cattle tank

41

that was covered with l-mm mesh ﬁberglass window screening between trials to prevent
colonization by other organisms. To maintain sufﬁciently large populations of the three
Chaoborus species, the cattle tank was restocked 24 hours prior to the start of each trial.
Chaoborus species abundance prior to the start of each feeding trial was determined by
sampling the entire water column of the cattle tank with 2 vertical tows using a 30 cm
diameter, 500-micron mesh net. For each trial, ﬁve randomly chosen bluegill from a pool
of 50 bluegill, ranging in size from 59-89 mm SL, were starved for 24 h and then placed
in the Chaoborus stocked tank. Bluegill were allowed to feed for 10 min to minimize the
effect of prey depletion. At the end of each trial, bluegill were removed, measured, and
stomach ﬂushed using deionized water from a 20 cc syringe. Chaoborus sampled pre-
trial and bluegill stomach contents were preserved in 95% ethanol, identiﬁed to species,
counted, and measured under 40X magniﬁcation using a digitizer tablet and soﬁware
(SigmaScan Pro Version 4.01; SPSS Inc., 1987). Bluegill preference for prey type i was
calculated using the Manly-Chesson index:

k

(11 = (di/ BI) / 2 (dj/ Ej)
j= 1

where i = l, 2, ..., k and where k is the number of prey types, d.- is the number of prey
type i in the diet summed across all ﬁve bluegill, and E, is the density of prey type i in the
cattle tank prior to ﬁsh addition (Chesson 1978, 1983, Manly 1974). Prey types that are
consumed in proportion to their abundance in the environment (i.e. no preference) have a,
= l/k, for this experiment k = 3. Prey types that are selected for have a, > l/k and prey
types selected against have a, < l/k.

Bluegill selectivity for the three species of Chaoborus was analyzed using the
mean 01 of eight trials by species versus l/k = 0.33 in a one-sample t-test. The difference
in Chaoborus size distributions in the environment versus bluegill diet was analyzed

using a two-sample Kolmogorov-Smirnov test.

42

To examine whether migratory behavior and the presence of a refuge from ﬁsh
predation would affect selectivity of the three Chaoborus species, a 5 mm mesh plastic
screen was placed at half the depth of the cattle tank one hour before sunset on the day
before each trial for a total of four trials. Chaoborus species abundance in the cattle tank
was measured prior to the addition of the refuge with 2 vertical tows. The abundance of
Chaoborus above the refuge (non-migratory) was measured the morning after refuge
addition with 2 vertical tows that sampled the water column above the refuge. For each
trial ﬁve bluegill were added to the cattle tank using the same methods as in the non-
refuge feeding trials but were allowed to feed for only 8 minutes. At the end of each trial
bluegill were removed, measured, and stomach ﬂushed using deionized water from a 20
cc syringe. Chaoborus sampled pre-trial and bluegill stomach contents were preserved in
95% ethanol, identiﬁed to species, counted, and measured under 40X magniﬁcation using
a digitizer tablet and software (SigmaScan Pro Version 4.01; SPSS Inc., 1987). Bluegill
preference for each prey type was calculated using the Manly-Chesson index.

Bluegill selectivity for the three species of Chaoborus was analyzed using the
mean a of four trials by species versus 1/k = 0.33 in a one sample t-test. The difference in
Chaoborus size distributions above the refuge versus bluegill diet and above the refuge
versus the entire cattle tank was analyzed using a two-sample Kolmogorov-Smirnov test.
The effect of the refuge was analyzed using the mean proportion of C haoborus by species
found above the refuge versus a random distribution of Chaoborus species in a one-
sample t-test. All analyses were performed using Systat 11.0 (SPSS Inc., 2004).

Results
Fish Gradient Experiments: Species sorting and colonization

The results of Experiment 2 are presented ﬁrst, as the larger numbers of
Chaoborus collected in this experiment provide the clearest picture of Chaoborus species
sorting along the bluegill density gradient. The results of Experiment 1 support the

patterns observed in Experiment 2 and are presented second.

43

In Experiment 2, the four C haoborus species showed dramatically different
responses to ﬁsh predation. At the start of the experiment (prior to ﬁsh introduction) C.
americanus was present at a high density in all the ponds and the other three C haoborus
species were absent (Fig. 4.1). This pattern is consistent with the observed species
distributions in nature (Garcia, Chap. 2). C. americanus remained abundant in ponds
without ﬁsh throughout the experiment, declined slightly in the medium ﬁsh biomass
ponds, and declined dramatically in the high ﬁsh biomass ponds (Fig. 4.1a). The
differential response between the ﬁsh biomass categories but overall general decline
through time was supported by the rmANOVA, which showed a signiﬁcant treatment and
time effect (p=0.04, F2,12=4.20, between subjects; p=.002, F2,24=7.94, within subjects time
effect, rmANOVA). The other three species of Chaoborus ﬁrst appeared in the ponds on
sampling day 14. C. punctipennis density increased through time (leveling off by day 84)
and this increase was most noticeable in the high ﬁsh biomass ponds (Fig. 4.1b; p=0.003,
F 2,12=9.82, between subjects, and p=0.003, F2,24=7.41, within subjects time effect,
rmANOVA). C. ﬂavicans density also increased through time in all of the ponds but there
were no signiﬁcant differences between ﬁsh biomass categories (Fig. 4.10; p>0.05, for all
within and between subjects ﬁsh biomass effects and time by biomass interactions,
rmANOVA). C. albatus density increased through time in the medium and high ﬁsh
biomass ponds but differences between ﬁsh biomass categories were not signiﬁcant (Fig.
4.1d; p>0.2, for all within and between subjects ﬁsh biomass effects and time by biomass
interactions, rmANOVA).

Looking at Chaoborus densities averaged over the last three sampling dates
showed that C. americanus density signiﬁcantly declined with ﬁnal ﬁsh biomass across
all ponds (Fig. 4.2a; r2=0.45, p=0.006). C. punctipennis and C. albatus showed the
opposite pattern, as both showed a strong, positive response to ﬁnal ﬁsh biomass (C.
punctipennis, Fig. 4.2b; r2=0.71, p<0.001; C. albatus, Fig. 4.2d; r2=0.39, p=0.013). Final

ﬁsh biomass had no signiﬁcant effect on C. ﬂavicans density although there was a slight

44

 

chm

vomm

wam

wwNm

moon

ﬁmvm

mwvm

enm—

N_b

5mm

Auxaa:=xn
gnu—ace—

 

whcm

_mMN

acm—

h_o_

m_b

hob

com

mom

mam

vom

mm—

vamniaox_
sic—ag::_

 

ON

N—

N—

3.?
=52: ha a

 

 

@—

 

m—

@—

 

 

 

 

m—

 

 

N—

 

O—

 

 

h—

 

 

V—

 

 

 

Aw7ﬁvm

 

 

Am

 

hmom_

om_¢

wmmw

vmmw

omhm

Oh_m

385:2: .2. in:

 

wmom

_ch

momm

wmmm

w—c

aw—

hv

@ 82oz .2. .3...—

 

oom—

ocm

cow

oo—

o~

33a .52.: .o u

 

 

w—

 

~—

 

 

 

 

 

@—

 

h—

 

m—

 

 

Azﬁnr—

 

 

2

@855 i: Ea 95 BE .23 .m 295% E n5 _ SEEK“ Q 5 cyan 352%». E23.» 5: .3 28¢

45

 

10

a) C. amen’canus b) C. pmctipermis

014

0.01 4

 

 

 

 

 

 

 

 

 

 

 

 

 

A
..J
g + Zero ﬁsh biomass
3‘ 0°01 ‘ -o— Medﬁshbfuomass *
a + High ﬁsh biomass
C
m om1 Y 1 Y V V 7
'0 10 _
g c) C. ﬂawmns d) C. albatus
K
O
Q 1 <
O
(U
Q
0 01

0.01 «

0001 «

0m1 r Y Y Y T Y 1 T
0 20 40 60 80 100 120 0 20 40 60 80 100 120
Sampling Day

Figure 4.1. Change in Chaoborus density (number per liter) through time in Experiment 2
grouped into three ﬁsh density categories. Time is the sampling day of the experiment (0-
112, May-September 2005). The ﬁlled circles are the means (i 1 s.e., n=4) for the zero
ﬁsh biomass ponds, the open circles are the means (i 1 s.e., n=4) for the medium ﬁsh
biomass (500-1250 g) ponds, and the ﬁlled triangles are the means (i: 1 s.e., n=7) for the
high ﬁsh biomass (3250-5850 g) ponds.

46

 

 

V a) C. americanus 5) C- PUMPOFWS

r2=0.447 r2=0.711 .
v P=0.006 P<0-001

01*

 

 

 

 

 

 

 

 

 

 

 

: 0.014
E
.E‘
a)
c
GJ 0001
U 1
g c) C. ﬂavicans d) C. albatus
8 r2=0.203 r2=0.390
8 P=0.092 P=0.013
~E on U 4
0
U A
...... D A ‘
0°11 ......... {3'6 ........................ ....... 0 g .
g [I] ........ 8t] ................ D ‘ ‘
W a 2;. a. .03. m a .03. .5... .50. .00.

Final Fish Biomass (9)

Figure 4.2. Chaoborus density (number per liter) in Experiment 2 averaged across the last
three sampling dates as a function of ﬁnal ﬁsh biomass. Lines are the ﬁtted linear
regressions for each species (solid lines are signiﬁcant, p<0.01) and each point represents

a pond.

47

negative trend (Fig. 4.20; r2=0.2, p=0.09). Final ﬁsh biomass also had a signiﬁcant
negative effect on the number of days C. americanus was present in the ponds (Fig. 4.3;
8:065, p<0.0001).

In Experiment 1, the Chaoborus species assemblage responded similarly to the
gradient in ﬁsh biomass as in Experiment 2, although the results were less striking due to
the lower number of Chaoborus individuals sampled. C. americanus was observed in
seven of the ten ponds at low densities prior to ﬁsh introduction when compared to the
densities of C. americanus in the ponds at the start of Experiment 2 (Figs. 4.1a and 4.4a).
These initial density differences are not surprising given that C. americanus had a full
year to colonize the ponds in Experiment 2 and only a couple weeks in Experiment 1. C.
americanus densities declined rapidly in the medium and high ﬁsh biomass ponds, but
increased in abundance in the ﬁshless ponds throughout the experiment (Fig. 4.4a:
p=0.01, F2,7=8.96, between subjects biomass effect, rmANOVA). Final C. americanus
abundances were similar in ﬁshless ponds in the two experiments. C. ﬂavicans was
present in three of the ten ponds prior to ﬁsh introduction and showed no signiﬁcant
changes in density through time or across ﬁsh treatments (Fig. 4.40; p>0.2, for all within
and between subjects ﬁsh biomass effects and time by biomass interactions, rmANOVA).
C. punctipennis ﬁrst appeared in the ponds on sampling day 10 and then increased
through time. As in Experimental 1, this increase was most noticeable in the high ﬁsh
biomass ponds, although there were no signiﬁcant differences (Fig. 4.4b; p>0.2, for all
within and between subjects ﬁsh biomass effects and time by biomass interactions,
rmANOVA). C. albatus densities increased through time only in the high ﬁsh biomass
ponds but there were no signiﬁcant differences between ﬁsh biomass categories (Fig.
4.4d; p>0.2, for all within and between subjects ﬁsh biomass effects and time by biomass

interactions, rmANOVA).

48

120

 

 

 

 

 

o a o o r2=0.65

100 . P<0.001
,_, 80 -
C
a:
(I)
9
a 604
‘9. 0
co
0 4o .

201

0 I I I T

0 2000 4000 6000 8000

Final Fish Biomass (9)

Figure 4.3. Number of days C. americanus was present in the water column in
Experiment 2 as a function of ﬁnal ﬁsh biomass. The solid line is the ﬁtted linear

regression (p<0.01) and each point represents a pond.

49

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1
la) C. americanus

0.1 <
A
d 0.01 -
11:
v
.é‘
m
c
3 0.0011
0)
0
Q
0
<0
5 0.1~

0.01 ‘

o-m1 V I f 1' j Y T y 1 r Y I I I
0 Z] 0 m m 100 120 10 0 Z) 40 m m 100 12) 14)
Sampling Day

Figure 4.4. Change in Chaoborus density (number per liter) through time in Experiment 1
grouped into three ﬁsh biomass categories. Time is the sampling day of the experiment
(0-123, May-September 2003). The ﬁlled circles are the means (i 1 s.e., n=3) for the zero
ﬁsh biomass ponds, the open circles are the means (i 1 s.e., n=2) for the medium ﬁsh
biomass (2300-3200 g) ponds, and the ﬁlled triangles are the means (3: 1 s.e., n=5) for the
high ﬁsh biomass (5700-12,050 g) ponds.

50

Looking at Chaoborus densities averaged over the last ﬁve sampling dates
showed that C. americanus was present only in the ponds without ﬁsh (Fig. 4.5a; r2=0.45,
p=0.03). C. punctipennis and C. albatus again showed the opposite pattern, as both
showed a positive response to ﬁnal ﬁsh biomass (C. punctipennis, Fig. 4.5b; r2=0.33,
p=0.09; C. albatus, Fig. 4.5d; r2=0.73, p=0.002), but ﬁnal ﬁsh biomass had no signiﬁcant
effect on C. flavicans density (Fig. 4.50; r2=O.O2, p=0.74). Thus, the combined results
from both experiments found that C. americanus showed an overall negative response to
increasing planktivore abundance, whereas C. punctipennis and C. albatus showed
positive responses, and there was no trend for C. ﬂavicans.

Prey Preference Trials: Non-refuge and refuge

In the absence of a refuge, bluegill tended to prefer C. americanus (Fig. 4.6a,
p=0.10, t7=1.87), whereas C. punctipennis was strongly selected against (Fig. 4.6a,
p=0.01, t7=-3.52), and there was no preference for or against C. ﬂavicans (Fig. 4.6a,
p=0.843, t7=-O.21). In ﬁve out of eight feeding trials, the size distribution of all
Chaoborus (C. punctipennis, C. ﬂavicans, and C. americanus) eaten by the bluegill was
signiﬁcantly larger than that found in the environment (Fig. 4.7a, p<0.06). When size-
selection was analyzed within species, C. americanus and C. punctipennis were larger in
the diet than in the environment (p<0.06) but this was not true for C. ﬂavicans.

The presence of a refuge did not alter the pattern of species selection (Fig. 4.6b)
and few C. ﬂavicans were found above the refuge indicating that they vertically migrated
(p=0.02, t3=-6.38). In all four trials, the size distribution of the Chaoborus assemblage (C.
punctipennis, C. ﬂavicans, and C. americanus) found above the refuge was smaller than
the size distribution found in the entire water column before adding the refuge (p<0.03);
indicating that the larger Chaoborus moved below the refuge. This result was largely due
to the inﬂuence of C. americanus. A comparison of the size distributions by species
showed that C. americanus was the only species whose mean size above the refuge was

smaller than in the absence of the refuge (p<0.01). As in the non-refuge experiment, ﬁsh

51

 

 

 

 

 

 

 

 

 

 

 

 

v a)C.amricanus b)C.punctipermis
r2=o.448 r2=o.326 '
01‘ P=0.085
o
. C
001- .
A o
..l
E 0001- o '
.é‘
m
c
0 0.0001 ﬁ
IO 1
‘é’ c) 0 ﬂavicansl d) c. albatus
o r2=0.015 r2=0.732
8 P=0.739 P=0.002
g ml .
0
D
l D A
0.01 U D
............................................................... E1 A ‘ 5
D o a 0 D A
0031 T 1 , w v u r . v v v v u v
0 2000 4000 SW 0000 10000 12000 14000 0 2000 4000 m am 10000 12000 1400

Final Fish Biomass (9)

Figure 4.5. Chaoborus density (number per liter) in Experiment 1 averaged across the last
ﬁve sampling dates as a function of ﬁnal ﬁsh biomass. Lines are the ﬁtted linear
regressions for each species (solid lines are signiﬁcant, p<0.01) and each point represents

a pond.

52

 

a)

0.6 -l l

0.4 < T

 

0.2 -

Fish Preference (Manly-Chesson a)

 

 

 

 

 

 

 

 

0.0 . . .
C. americanus C. punctipennis C. ﬂavicans

 

b)

 

 

 

Fish Preference (Manly-Chesson a)

to

 

 

 

 

 

 

 

 

 

C. americanus C. punctipennis C. ﬂavicans

Figure 4.6. Average preference (Manly-Chesson 0. i l s.e.) of Bluegill for three species
of Chaoborus in a) 8 non-refuge feeding trials and b) 4 refuge feeding trials. The line

indicates no preference.

53

were positively size selective. The size distribution of C haoborus eaten was larger than
the size distribution of Chaoborus found above the refuge (and therefore available to the
ﬁsh) (Fig. 4.7b, p<0.06). This result was signiﬁcant for C. americanus and C.
punctipennis (p<0.05), but not for C. ﬂavicans.

Discussion

It is widely appreciated that the presence of predators can affect the abundance
and distribution of prey species (Kerfoot and Sih 1987, Zaret 1980). However, how
variation in predator abundance may affect prey species sorting and colonization is less
well understood. In this study, I show that manipulating ﬁsh density can lead to clear
species sorting within an assemblage of four Chaoborus species, and that this response is
consistent with distributional patterns observed in nature. Turner and Mittelbach (1990)
report similar shifts in the Chaoborus species assemblage when they added 400 bluegill
to one of the KBS experimental ponds. Prior to ﬁsh addition, C. americanus and C.
ﬂavicans were present in the pond and C. americanus made up 95% of the C haoborus
population. However, within a week of adding ﬁsh, these species disappeared. Sometime
later in the experiment, C. punctipennis (>95% of the C haoborus population) and C.
albatus appeared in the pond. Turner and Mittelbach (1990) were unable to document
the time course of species replacement because their daytime sampling missed censusing
the vertically migrating C. punctipennis and C. albatus when they ﬁrst appeared in the
pond. An important observation from the current study (which employs a regression
design as opposed to simple ﬁsh presence/absence) is that the pattern in the Chaoborus
species abundance depends quantitatively on ﬁsh density.

Many observational studies have found that C. americanus only occurs in ﬁshless
habitats (Garcia, Chap. 2, Wissel et al. 2003). von Ende (1979) did observe one instance
of a lake with only a single species of ﬁsh that also contained C. americanus. He stated
that, “. . .although the exclusion of C. americanus by ﬁsh appears to be a fairly general

phenomenon, the intensity of predation depends on the interaction of the characteristics

54

 

 

- Environment
122:“) Diet

0.16 4

 

 

 

 

0.14 -
0.12 -i
0.10 o
0.08 ..

0.06 4

Proportion of Chaoborus

0.04 '1

0.02 -1

 

 

 

 

 

 

 

 

 

0.00

 

14 16

 

b)

 

0.08 -*

0.06 1

Proportion of Chaoborus

0.04 -‘

 

 

0.02 -

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0.00 u

2 4 6 8 10 14 16

Prey Length (mm)

Figure 4.7. Frequency distribution of Chaoborus lengths (includes C. punctipennis, C.
ﬂavicans, and C. americanus) found in the environment (gray bars) and ﬁsh diet (black

bars) in a) non-refuge feeding trials (n=8 trials) b) refuge feeding trials (n=4 trials).

55

of the lake, the C. americanus population, and the ﬁsh species.” In my experimental
gradients, C. americanus was quickly eliminated in the high ﬁsh biomass ponds but in
Experiment 2, C. americanus was able to persist in ponds with medium ﬁsh biomass
(Figs. 4.1a, 4.3, and 4.5a). This result was not observed in the ﬁrst experiment, likely due
to the fact that C. americanus densities, prior to ﬁsh addition, were much lower and
therefore more vulnerable to local extinction.

Unlike C. americanus, most studies ﬁnd that C. punctipennis commonly coexists
with ﬁsh (von Ende 1979, Ramcharan et al. 2001, Carter et al. 1980, Wissel et al. 2003,
Garcia, Chap. 2). Not only was C. punctipennis found in all the experimental ponds with
ﬁsh, but recruitment of C. punctipennis was particularly favored in the high ﬁsh biomass
ponds (Figs. 4.1b and 4.5b). This result is supported by Wissel et al. (2003), where they
found that C. punctipennis was positively correlated with abundance and presence of ﬁsh.
They also observed that C. punctipennis tended to reach higher densities in shallow lakes
which makes sense because shallow lakes tend to support higher levels of ﬁsh
planktivory (Tessier and Woodruff 2002). Ramcharan et al. (2001) also found that C.
punctipennis was positively associated with high planktivory.

The increase in C. punctipennis and C. albatus density with increasing ﬁnal ﬁsh
biomass may be explained by the elimination of C. americanus in those systems (Fig.
4.2a and b). Little is known about C. albatus because it is a rare species (Garcia, Chap.
2), but von Ende (1979) observed that late instars of C. americanus will prey on C.
punctipennis and he posits this as the mechanism excluding C. punctipennis from systems
where C. americanus is present. Here I propose another possible mechanism, where
recruitment of C. punctipennis and possibly C. albatus is favored in environments with
ﬁsh due to the indirect effect of ﬁsh on their shared zooplankton prey. Many studies
have found that in the presence of ﬁsh the zooplankton community shifts toward greater
dominance by small-bodied zooplankton (Vanni 1987, Brooks and Dodson 1965). Since

C haoborus are gape-limited predators and C. punctipennis and C. albatus are the smallest

56

of the four species found in this region, recruitment of these two species would be
expected to be favored in environments with abundant small sized zooplankton. When I
examined mean zooplankton size in ponds grouped into the three ﬁsh biomass categories
I found a signiﬁcant negative effect of high ﬁsh biomass on mean zooplankton size
(p<0.05, one-way ANOVA). Further, the lack of coexistence between C. americanus and
C. punctipennis may be a result of a combination of the mechanism proposed here and
that of von Ende (I979).

The neutral response of C. ﬂavicans to my experimental bluegill gradient is also
supported in the literature and from a survey of local lakes and ponds (Pope et a1 1973,
Gonzalez and Tessier 1997, Garcia, Chap.2). Wissel et a1 (2003) found that C. ﬂavicans
presence was negatively correlated to lake area, positively correlated with DOC levels
and had no relationship with the presence or absence of ﬁsh. In my survey I found that C.
ﬂavicans was found in both ponds without ﬁsh and often in greater abundance in lakes
with ﬁsh. Berendonk et al. (2003) found that C. ﬂavicans is a comparatively “new” lake
lifestyle species and this could be related to its neutral response. They state that at least
three shifts have occurred between pond and lake lifestyles for Chaoborus species and
that this shift is often accompanied by a decrease in larval body size, and associated with
the evolution of DVM behavior. Although C. ﬂavicans does exhibit DVM behavior this
may be a plastic response to ﬁsh chemicals because C. ﬂavicans has shown a marked
decrease in migratory behavior when not exposed to ﬁsh chemicals (McQueen et al.
1999, Tjossem 1990). It may be that coexistence with ﬁsh is due to C. ﬂavicans’s
transparency and DVM behavior. In addition, C. ﬂavicans’s potential ability to modify its
migration behavior and its large size may enable it to persist in the larger prey size
environments of ﬁshless habitats but also be why it is not positively associated with high
ﬁsh biomass like C. punctipennis.

While a number of studies report Chaoborus in bluegill diets (e.g., Mittelbach

1981 , Turner and Mittelbach 1990), no one has speciﬁcally examined foraging preference

57

by bluegill (or other ﬁsh species) for different species of Chaoborus. The prey
preference experiments reported here indicate that larval body size and DVM behavior
are important traits determining the vulnerability of different Chaoborus species to
bluegill, and these traits should therefore inﬂuence the distribution of Chaoborus species
across the ﬁsh density gradient (Berendonk et al. 2003). The observation that large (III
and IV instar) C. americanus may have vertically migrated while smaller (I and II instar)
C. americanus did not was a surprising result and has not been found in previous studies.
This may be another possible mechanism by which C. americanus was able to persist in
the medium ﬁsh biomass ponds. Another surprising result was the lack of preference for
C. ﬂavicans in the non-refuge feeding trials because both C. americanus and C. ﬂavicans
should be preferred because of their large body size relative to the other C haoborus
species. The lack of preference for C. ﬂavicans may be understandable in the context of
the refuge feeding trials since C. ﬂavicans exhibited DVM behavior in response to the
refuge. Thus, selection for C. americanus and not C. ﬂavicans may be more to do with
the combination of pigmentation and large size in C. americanus (Stenson 1980). Duffy
et al. (2005) found strong evidence of bluegill preferring Daphm'a that were pigmented
due to infection by a parasite over similarly sized Daphnia that were not infected and
thus not pigmented.

The experiments presented here illustrate that the processes responsible for the
distributional pattern of Chaoborus species in nature is more involved than the simple
presence or absence of a ﬁsh predator. They also suggest that C haoborus species traits
such as DVM behavior, size and pigmentation, are functionally related to the
determinants of species distributions and development and maintenance of community
structure across the gradient. Aquatic systems vary in size, depth and the presence and
abundance of piscivores and this variation is related to differences in the intensity of ﬁsh
planktivory (Tessier and Woodruff 2002). Future work on the effect of gradients in the

density of competitors and predators on species sorting are needed because such datasets

58

are readily useful for incorporation into simulation models of food web dynamics
(Cottingham et al. 2005). This will increase the applicability of such models to natural
systems and enhance our understanding of factors that are important to the development

and maintenance of community structure.

59

LITERATURE CITED

Abjornsson, K., C. Brdnmark, and L. Hansson. 2002. The relative importance of lethal
and non-lethal effects of ﬁsh on insect colonisation of ponds. Freshwater biology
47:1489-1495.

Binckley, C. A. and W. J. Resetarits Jr. 2003. Functional equivalence of non-lethal
effects: generalized ﬁsh avoidance determines distribution of gray treefrog, Hyla
chrysoscelis, larvae. Oikos 1022623-629.

Berendonk, T. U. 1999. Inﬂuence of ﬁsh kairomones on the ovipositing behavior of
Chaoborus imagines. Limnology and Oceanography 44(2):454-458.

Berendonk, T. U., and M. B. Bonsall. 2002. The phantom midge and a comparison of
metapopulation structures: Ecology 83:1 16—129.

Berendonk, T. U., T. G. Barraclough, and J. C. Barraclough. 2003. Phylogenetics of pond
and lake lifestyles in Chaoborus midge larvae. Evolution 57(9):2173-2178.

Blaustein, L. 1999. Oviposition site selection in response to risk of predation: evidence
from aquatic habitats and consequences for population dynamics and community
structure. In: Wasser SP (ed) Evolutionary theory and processes: modern
perspectives. Kluwer, Dordrecht, pp. 441-456.

Blaustein, L., M. Kiﬂawi, A. Eitam, M. Mangel, J .E. Cohen. 2004. Oviposition habitat
selection in response to risk of predation in temporary pools: mode of detection
and consistency across experimental venue. Oecologica 138:300-305.

Borkent, A. 1979. Systematics and bionomics of the species of the Subgenus
Schadonophasma Chaoborus Chaoboridae Diptera. Quaestiones Entomologicae
15: 122—256.

Borkent, A. 1981. The distribution and habitat preferences of the Chaoboridae

Culicomorpha Diptera of the Holarctic region: Canadian Journal of Zoology
59:122—133.

Brooks, J. L. and S. I. Dodson 1965. Predation, body size, and composition of plankton.
Science 150228-35.

Carter, J. C. H., M. J. Dadswell, J. C. Roff, and W. G. Sprules. 1980. Distribution and
zoogeography of planktonic crustaceans and dipterans in glaciated eastern North
America. Canadian Journal for Zoology 58:1355-1387.

Carter, J. C. H., J. K. Kwik. 1977. Instar succession, vertical distribution, and

interspeciﬁc competition among four species of Chaoborus. Journal of Fisheries
Research Board of Canada 34:113-118.

60

Chesson, J. 1978. Measuring preference in selective predation. Ecology 59(2):211-215.

Chesson, J. 1983. The estimation and analysis of preference and its relationship to
foraging models. Ecology 64(5):1297-1304.

Cook, E. F. 1956. The Nearctic Chaoborinae (Diptera: Culicidae): Technical Bulletin of
University Minnesota Agricultural Experiment Station 21821—102.

Cottenie, K. 2005. Integrating environmental and spatial processes in ecological
community dynamics. Ecology Letters 8(11):1175-1182.

Cottingham, K. L., J. T. Lennon, and B. L. Brown. 2005. Knowing when to draw the line:
designing more informative ecological experiments. Frontiers in Ecology and the
Environment 3(3): 145-1 52.

Darcy-Hall, T. L. 2006. Oecologia (in press)

Dawidowicz, P., J. Pijanowska, and K. Ciechomski. 1990. Vertical migration of
Chaoborus larvae is induced by the presence of ﬁsh. Limnology and
Oceanography 35(7): 163 1-1637.

Duffy, M.A., S.R. Hall, A.J. Tessier and M. Huebner. 2005. Selective predators and their
parasitized prey: Are epidemics in zooplankton under top-down control?
Limnology and Oceanography 50(2):412-420.

Gliwicz, Z. M. and J. Pijanowska. 1989. The role of predation in zooplankton succession.
In Sommer, U. (ed.), Plankton Ecology: Succession in Plankton Communities.
Springer-Verlag, New York, pp.253-295.

Gonzalez, M. J. 1998. Spatial segregation between rotifers and cladocerans mediated by
Chaoborus. Hydrobiologia 387/388z427-436.

Gonzalez, M. J. and A. J. Tessier. 1997. Habitat segregation and interactive effects of
multiple predators on a prey assemblage. Freshwater Biology 38:179-191.

Kerfoot, W. C ., and A. Sih. 1987. Predation: Direct and indirect impacts on aquatic
communities: University Press of New England, Hanover, NH, USA.

Lamontagne, S., D. B. Donald and D. W. Schindler. 1994. The distribution of four
Chaoborus species (Diptera: Chaoboridae) along an elevational gradient in
Canadian Rocky Mountain lakes. Canadian Journal of Zoology 72: 1531-1 537.

Leibold, MA. 1996. A graphical model of keystone predators in food webs: trophic

regulation of abundance, incidence and diversity patterns in communities.
American Naturalist 147:784-812.

61

MacArthur, RH. & Wilson, EC. 1967. The Theory of Island Biogeography. Princeton
University Press, Princeton, NJ, USA.

Manly, B.F.J. 1974. A model for certain types of selection experiments. Biometrics
30:281-294.

McCune, G. and J. B. Grace. 2002. Analysis of Ecological Communities. MjM software
Design, Gleneden Beach, Oregon, USA.

McCune, G. and M. J. Mefford. 1999. PC-ORD. Multivariate analysis of ecological data,
v4.0. MjM Sotware Design, Gleneden Beach, Oregon, USA.

McPeek, M.A. 1989. Differential dispersal tendencies among Enallagma damselﬂies
(Odonata) inhabiting different habitats. Oikos 56(2):187-195.

McQueen D.J., C. W. Ramcharan, E. Demers, N. D. Yan, L. M. Conforti, and A. Perez-
Fuentetaj a. 1999. Chaoborus behavioral responses to changes in ﬁsh density.
Archiv Hydrobiolgie 145(2): 165-179.

Mittelbach, G. G. 1981. Foraging efﬁciency and body size: a study of optimal diet and
habitat use by bluegills. Ecology 62:1370-1386.

Mittelbach, G. G. 1984. Predation and resource partitioning in two sunﬁshes
(Centrarchidae). Ecology 65 :499-5 13.

Mittelbach, G. G. and P. L. Chesson. 1987. Predation risk: indirect effects on ﬁsh
populations, p.237-289 in W. C. Kerfoot and A. Sih, (eds.), Predation: direct and

indirect impacts on aquatic communities. University Press of New England,
Hanover, NH, USA.

Mittelbach, G. G. 1988. Competition among refuging sunﬁshes and effects of ﬁsh density
on littoral zone invertebrates. Ecology 69(3):614-623.

Mittelbach, G. G. and C. W. Osenberg. 1993. Stage-structured interactions in bluegill:
consequences of adult resource variation. Ecology 74:2381-2394.

Mittelbach, G. G., A. M. Turner, D. J. Hall, J. E. Rettig, and C. W. Osenberg. 1995.
Perturbation and resilience: a long-term, whole-lake study of predator extinction
and reintroduction. Ecology 76:2347-2360.

Mittelbach, G. G. 2002. Fish foraging and habitat choice: a theoretical perspective, pgs.
251-266 in: P. J. B. Hart and J. D. Reynolds (eds.), Handbook of Fish and
Fisheries, Blackwell Science.

Mittelbach, G. G., E. A. Garcia, and Y. Taniguchi. 2006. Fish reintroductions reveal
smooth transitions between lake community states. Ecology 87(2):312-318.

62

Moore, M. V., 1986. Method for culturing the phantom midge, Chaoborus (Diptera:
Chaoboridae), in a laboratory. Aquaculture 56:307-316.

Moore, M. V., N. D. Yan, and T. Pawson. 1994, Omnivory of the larval phantom midge
(Chaoborus spp.) and its potential signiﬁcance for freshwater planktonic food
webs: Canadian Journal of Zoology 72:2055-2065.

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

Pope, G. F., J. C. H. Carter, and G. Power. 1973. The inﬂuence of ﬁsh on the distribution
of Chaoborus spp. (Diptera) and density of larvae in the Matamek River system,
Quebec. Transactions of the American Fisheries Society 102:707-714.

Ramcharan, C. W., N. D. Yan, D. J. McQueen, A. Pérez-Fuentetaja, E. Demers, and J. A.
Rusak. 2001. Complex responses of Chaoborus to changes in ﬁsh populations.
Archiv fur Hydrobiologie Special Issue Advances in Limnology 56281-100.

Resetarits, W .J., Jr. 2001. Colonization under threat of predation: non-lethal effects of
ﬁsh on aquatic beetles, T rapisternus lateralis (Coleoptera: Hydrophilidae).
Oecologia 129:155-160.

Rieger, J. F ., C. A. Binckley, and W. J. Resetarits, Jr. 2004. Larval performance and
oviposition site preference along a predation gradient. Ecology 85(8):2094-2099.

Smther, O. A. 1970. Nearctic and Palearctic Chaoborus, Diptera, Chaoboridae. Fisheries
Research Board of Canada Bulletin 57.

Saether, O. A. 1972. Chaoboridae, Das Zooplankton der Binnengewasser: Die
Binnengewasser. 26:257-281.

Sather, O. A. 1997. Diptera Chaoboridae, Phantom Midges, pgs. 149-161 in A. N.
Nilsson, (ed). Aquatic Insects of North Europe-A Taxonomic Handbook Vol. 2.
Apollo Books, Stenstrup, Denmark.

Stenson, J. A. E. 1980. Predation pressure from ﬁsh and two Chaoborus species as
related to their visibility. In: Evolution and Ecology of Zooplankton
Communities, American Society Limnology Oceanography Special Symposium 3
(Ed. W.C. Kerfoot), pp. 618-622. University Press New England, Hanover, NH,
USA.

Sterner, R. W. 1989. The role of grazers in phytoplankton succession. In: Sommer, U.
(ed.), Plankton ecology: succession in plankton communities. Springer-Verlag,
New York, pp. 107-169.

Stoks, R. and M. A. McPeek. 2003. Predators and life histories shape Lestes damselﬂy

63

assemblages along a freshwater habitat gradient. Ecology 84(6):1576-1 587.

Swift M. C., and A. Y. F edorenko. 1975. Some aspects of prey capture by Chaoborus
larvae. Limnology and Oceanography 20:418-425.

Tessier, A. J ., and P. Woodruff. 2002. Cryptic trophic cascade along a gradient of lake
size. Ecology 83(5):1263-1270.

Tilman, D. 1982. Resource Competition and Community Structure. Princeton University
Press, Princeton, NJ, USA.

Tjossem, S. F. 1990. Effects of ﬁsh chemical cues on vertical migration behavior of
Chaoborus. Limnology and Oceanography 35: 1456-1468.

Turner, A. M., and G. G. Mittelbach. 1990. Predator avoidance and community structure:
Interactions among piscivores, planktivores, and plankton: Ecology 71 :2241-
2254.

Urban, M. C. 2004. Disturbance heterogeneity determines freshwater metacommunity
structure. Ecology 85(11):2971-2978.

Vanni, M. J. 1987. Effects of nutrients and zooplankton size on the structure of a
phytoplankton community. Ecology 68:624-635.

von Ende, C. N. 1979. Fish predation, interspeciﬁc predation, and the distribution of two
Chaoborus species. Ecology 60: 119-128.

von Ende, C. N. 1982. Phenology of four Chaoborus species. Enviromnental Entomology
11:9—16.

Welschmeyer, N. A. 1994. Fluorometric analysis of chlorophyll a in the presence of
chlorophyll b and phaeopigments. Limnology Oceanography 39:1985-1992.

Werner, E. E. and D. J. Hall. 1988. Ontogenetic habitat shift in bluegill: the foraging rate-
predation risk trade-off. Ecology 69:1352-1366.

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: 1 540-1 548.

Werner, E. E. and M. A. McPeek. 1994. Direct and indirect effects of predators on two
anuran species along an environmental gradient. Ecology 75(5):]368-1382.

Wetzel, R. G. and Likens, G. E. 1991. Limnological Analyses. 2nd edition, Springer-
Verlag, New York, pp. 89-92.

Wissel, B., N. D. Yan and C. W. Ramcharan. (2003) Predation and refugia: implications

64

for Chaoborus abundance and species composition. Freshwater
Biology 48(8): 1421-143 1.

Zaret, T. M. 1980. Predation and freshwater communities. Yale University Press, New
Haven, CT, USA.

65

 

 

 

IIIIIIIIIIIII UNIVERSITY lBNARIES

Ill/11111111ill/ll!!!”Will/11111Hill/ll!

3 1293 02845 3110