lltHllWlWWl

fl

WHHUHWIHI)l

  

 

W

\

 

    

LIBRARY
Michigan State
University

This is to certify that the
dissertation entitled

COPEPODOLOGY IN ALPINE LAKES: LIMITATIONS
TO RECOVERY OF HESPERODIAPTOMUS
SHOSHONE AFTER EXOTIC FISH ERADICATION

presented by

ANDREW M. KRAMER

has been accepted towards fulfillment
of the requirements for the

PhD. degree in Fisheries and Wildlife and
Program in Ecology,
Evolutionary Biology and
Behavior
./\

 

 

[1% W

Major Professor’s Signature

fivrcrr [.7A a dd?

Date

MSU is an affinnative-action, equal-opportunity employer

 

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.

 

DAIEDUE

DAIEDUE

DATEDUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 pa/CIRC/DateDuelndd-p.1

 

 

COPEPODOLOGY IN ALPINE LAKES: LIMITATIONS TO RECOVERY OF
HESPERODIAPT OM US SHOSHONE AFTER EXOTIC FISH ERADICATION

By

Andrew M. Kramer

A DISSERTATION

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

Doctor of Philosophy

Department of Fisheries and Wildlife and Program in Ecology, Evolutionary Biology and
Behavior

2007

ABSTRACT

COPEPODOLOGY IN ALPINE LAKES: LIMITATIONS TO RECOVERY OF
HESPERODIAPT 0M US SHOSHONE AFTER EXOTIC FISH ERADICATION

By

Andrew M. Kramer

In Chapter 1, I took advantage of fish introduction and removal in alpine lakes to
experimentally test for the Allee effect at the whole-ecosystem scale. We conducted a
multi-lake experiment in which the copepod Hesperodiaptomus Shoshone was stocked
into lakes from which it had been extirpated by the introduction of non-native fish. The
stocked densities bracketed a hypothesized critical density (0.5-5m'3). Successful
recovery by the copepod was observed in only the lake that was stocked at the highest
density. Copepods stocked into small cages at high density survived and reproduced at
rates comparable to natural populations, confirming that the lakes were suitable habitat
for this species. In support of mate limitation as the mechanism underlying recovery
failure, I found a significant positive relationship between mating success and density
across experimental and natural H. Shoshone populations. Further evidence for Allee
effects was obtained from a mesocosm experiment with a related species,
Skistodiaptomus pallidus. Together, these lines of evidence support the importance of the
Allee effect to population recovery of H. Shoshone in the Sierra Nevada, and to diaptomid
copepods in general.

In Chapter 2, I examine the possibility that the Allee effect influences population
genetics. By setting a lower limit on population size, Allee effects can constrain the loss

of genetic variability due to genetic drift. In H. Shoshone, models, surveys and

experimental data suggest populations at densities less than 0.5 - 5 individuals/m3 are
unable to persist due to mate limitation. Combining this range of critical density with
reproductive characteristics and the distribution of habitat sizes in nature, I estimated
minimum effective population sizes for H. Shoshone. Our calculations suggest that >90%
of H. Shoshone populations in the Sierra Nevada are resistant to significant changes in
heterozygosity or genetic distance, and 70-75% of populations will lose <10% of allelic
richness, during bottlenecks or founder events. By setting a lower limit on population
size, Allee effects constrain the loss of genetic variability due to genetic drift.

In Chapter 3, I combine 3-dimensiona1 video analysis from laboratory
experiments with life history data to improve the precision and accuracy of existing
estimates of H. Shoshone critical density, and to increase the general understanding of
copepod mating behavior. I also investigate the impact of temperature variation on
mating behavior. I estimate H. Shoshone critical density to be 0.56 —— 1.3 rn'3 and to be
highly dependent on body size, primarily due to changes in swimming speed. H.
Shoshone swimming speed increased >25% as temperature increased from 5°C to 16°C.
The corresponding decrease in critical density was roughly equivalent to the decrease
resulting from the six-fold variation in net reproductive rate observed for large-bodied
alpine copepod species. The average swimming speeds measured for H. Shoshone (1.7 -
2.4 cm/s) are dramatically faster than those previously reported for similar sized calanoid
copepods. Rapid swimming and the ability to follow pheromone trails greatly improve
the ability of H. Shoshone to find mates, and the relationship between temperature and
critical density suggests that recovery or colonization events may be more likely to

succeed in warmer lakes and/or warmer years.

ACKNOWLEDGEMENTS

I need to thank my advisor, Orlando Samelle, for his patience and encouragement
and for being such an excellent example of quality research and mentoring. I want to
thank Roland Knapp, without him this research would have been impossible both
scientifically and practically. Thanks to Jeannette Yen for sharing her facilities,
expertise, and enthusiasm. I also want to thank my committee members, Jim Bence, Kim
Scribner, Gary Mittelbach, and Scott McNaught for there assistance in planning this
research and their comments on these papers. Additional thanks go to Blair Wilson, Greg
Goldsmith, Chris Brownfield, Trip Armstrong, Jodi Garton, Chris Archer, Jake Nicholl,
Matt Hegeman, Ericka Hegeman, Katie Armstrong and the numerous others who helped
me in the backcountry, to Dan Dawson, Sandi Roll and the staff of the Sierra Nevada
Aquatic Research Laboratory, to Rachel Lasley, Megan Heaphy, Jennifer Sehn and the
members of Jeannette Yen’s lab, and to Ryan Lockwood, Paul Weidman and Brian
Parker for copepods. I want to acknowledge everyone in the Fisheries and Wildlife
Department, specifically my labmates: Alan Wilson, Lesley Knoll, Geoff Horst, Kendra
Cheruvelil, Kevin Pangle, Stacy Nelson, and Sherry Martin. I also need to acknowledge
Nicole Lamp for help not only with the substance of this document but for her love and
for constant support during all stages of this research. This work was funded by a
National Science Foundation Graduate Research Fellowship, Michigan State University
Distinguished Student Graduate Fellowship, National Science Foundation Grants DEB-
9629473, DEB-0075509, an REU supplement, VESR Graduate Student grant, and a

Sigma-Xi Grant-in-aid of Research.

iv

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................ vii
LIST OF FIGURES .......................................................................................................... ix
INTRODUCTION ........................................................................................................... 1

CHAPTER 1: ALLEE EFFECT LIMITS COLONIZATION SUCCESS OF SEXUALLY
REPRODUCING ZOOPLANKTON

ABSTRACT .......................................................................................................... 5
INTRODUCTION ................................................................................................ 7
METHODS ........................................................................................................... 9
Whole-lake experiment ............................................................................. 10
Assessment of mate limitation .................................................................. 16
C aging experiment .................................................................................... 17
Mesocosm experiment .............................................................................. 18
RESULTS ............................................................................................................. 20
DISCUSSION ....................................................................................................... 26
ACKNOWLEDGEMENTS .................................................................................. 33
LITERATURE CITED ......................................................................................... 34

CHAPTER 2: LIMITS TO GENETIC BOTTLENECKS AND FOUNDER EVENTS
IMPOSED BY THE ALLEE EFFECT

ABSTRACT .......................................................................................................... 38
INTRODUCTION ................................................................................................ 39
METHODS ........................................................................................................... 42
RESULTS ............................................................................................................. 45
DISCUSSION ....................................................................................................... 52
ACKNOWLEDGEMENTS .................................................................................. 57
LITERATURE CITED ......................................................................................... 58

CHAPTER 3: THE EFFECT OF MATING BEHAVIOR AND TEMPERATURE
VARIATION ON THE CRITICAL POPULATION DENSITY OF A FRESHWATER
COPEPOD

ABSTRACT .......................................................................................................... 62
INTRODUCTION ................................................................................................ 63
METHODS ........................................................................................................... 65
Animal collection and maintenance .......................................................... 65
Temperature experiments .......................................................................... 65
Encounter model and estimate of critical density ..................................... 68
RESULTS ............................................................................................................. 70
DISCUSSION ....................................................................................................... 73
ACKNOWLEDGMENTS .................................................................................... 79
LITERATURE CITED ......................................................................................... 80
APPENDIX A ................................................................................................................... 83
SAMPLE AREA ................................................................................................... 83
METHODS ........................................................................................................... 86
LIFE HISTORY .................................................................................................... 86
Hatching .................................................................................................... 86
Development ............................................................................................. 87

Body Size ................................................................................................... 87

Sex ratio .................................................................................................... 88
Reproductive rate ...................................................................................... 9O
Mortality ................................................................................................... 9O
LITERATURE CITED ......................................................................................... 91
LITERATURE CITED ..................................................................................................... 92

vi

LIST OF TABLES

CHAPTER 1: ALLEE EFFECT LIMITS COLONIZATION SUCCESS OF SEXUALLY
REPRODUCING ZOOPLANKTON

Table 1. Lake morphometry and fish history for the 6 experimental and 2 source
lakes ...................................................................................................................... 12

Table2. Experimental design, detailing the target stocking density for each
experimental lake, the number of copepods needed to attain that density, the
source of the stocked animals, and the logistical details of the stocking
event ...................................................................................................................... 13

Table 3. The volume sampled in each lake each year. The total number of
copepods collected in that volume is in parentheses ............................................ 15

Table 4. Reproductive and survival data from the caging experiment, the source of
the animals used in the caging experiment (Dissertation Lake), and the
successfully reintroduced population (Square Lake). Clutch size is average
and standard deviation of eggs per egg bundle over the course of the
experiment (cages) or the summer of 2004 (Dissertation and Square Lake).

# clutches/ female is the number of clutches per reproductive (gravid and/or
egg-bearing) female for 2 cages per lake after 2 weeks in the cages (cages)

or for the single closest sampling date (lakes). All sampling took place

between 7/19/2004-7/21/2004. Mortality is the average daily mortality
calculated from the regression of1n(density) vs. time. A-Mortality is mean

of 4 cages sampled on 3 dates from 7/11/2004-8/8/2004. B- Dissertation

Lake mortality is from 3 sampling dates from 7/29/2004 — 8/22/2004,

equivalent data is not available for same time period as the cages. C — Square
Lake mortality data is from 4 sample dates .......................................................... 25

CHAPTER 3: THE EFFECT OF MATING BEHAVIOR AND TEMPERATURE
VARIATION ON THE CRITICAL POPULATION DENSITY OF A FRESHWATER
COPEPOD

Table 1. The conditions and results of mating trials. Density and sex ratio varied
based upon mortality. The same individuals were used in trial 1 and 2,
except for replacements due to mortality between trials. Observed mating
attempts are the total seen in the recording of the 4 hour trial and analyzable
encounters are the subset from which complete 3D positions were
obtainable .............................................................................................................. 67

APPENDIX A

Table 1. Table 1: Lakes sampled, with lake ID (when known), name (when in
quotes the name is not found of topographical maps), morphological data,
stocking history(S-F=stocked, fish present, NS=never stocked, S—

vii

F L=stocked previously, now fishless, S-F L, Exp=Now fishless,

experimentally reintroduced), the number and years each lake was sampled,
additional samples which provided no information to this analysis are not
included. (sampling differs for experimentally populations, see Methods) .......... 84

viii

LIST OF FIGURES

CHAPTER 1: ALLEE EFFECT LIMITS C OLONIZATION SUCCESS OF SEXUALLY
REPRODUCIN G ZOOPLANKTON

Figure 1. Map of experimental lakes. A) Humphreys Basin is on the west side of
the Sierra Crest in the Sierra National Forest, a) Dissertation Lake, source of
H. Shoshone used in reintroduction, 1) No Good Lake, 2) Cony Lake, 3)
Square Lake, 4) Knob Lake. B) Morgan Creek Basin is located on the east
side of the crest in the Inyo National Forest is 12 km north of Humphreys
Basin, b) Spire Lake, source of H. Shoshone used in reintroduction, 5) Finch
Lake, 6) Middle Morgan Lake .............................................................................. 11

Figure 2. Target stocking density and observed density in the weeks following
(2003) in the experimental lakes (Ioglo scale) ...................................................... 21

Figure 3. Observed densities in the experimental lakes (logio scale) immediately
following reintroduction (2003) and in the following years. Lakes are
identified by target density and are in the same order as Fig. 2. Detection
limit < 0.005 m' unless noted otherwise .............................................................. 23

Figure 4. Proportion of reproductive females bearing eggs vs. density of adult

copepods in reintroduced populations (A) and natural populations (0).

Filled triangles (A) represent Square Lake, the reintroduction lake where

the population recovered. The line is the logistic regression function:

WithEggs = (eA-l.184 + 0.74 * AdultDensity) / (1 + (eA-1.184 + 0.74 *

AdultDensity)) (p < 0.0005) ................................................................................. 24
Figure 5. Daily population growth rate vs. introduced density of S. pallidus in 1000

L mesocosms. The line is the linear regression function: y = 0.003*x +

0.092, p = 0.025, R2 = 0.13, n = 38. Final density was undetectable in four

cases resulting in negative growth rate (see Methods). Experiment lasted for

4 weeks .................................................................................................................. 27

CHAPTER 2: LIMITS TO GENETIC BOTTLENECKS AND FOUNDER EVENTS
IMPOSED BY THE ALLEE EFFECT

Figure 1. The quantile distribution of lake volume for a representative sample of
H. Shoshone populations in the Sierra Nevada (n=169), and the
corresponding minimum effective population size for a critical density of
0.5/m3 .................................................................................................................... 46
Figure 2. The expected ratio of post-bottleneck heterozygosity (HB) to initial
heterozygosity (H,) as impacted by habitat size and duration of bottleneck
for (A) minimum density of 0.5 individuals/m3; (B) minimum density of 5
individuals/m3. Bottleneck duration is the number of generations population
is at minimum effective population size. A habitat of volume 64 m3 is the

largest habitat in which effective population size can decline to one male

and one female when critical density = 0.5/m3 ..................................................... 47
Figure 3. The proportion of original allelic richness retained after population

bottleneck as determined by habitat size and bottleneck duration for (A)

minimum density of 0.5 individuals/m3; (B) minimum density of 5

individuals/m3. Bottleneck duration is the number of generations population

is at minimum effective population size. Data was produced using the

simulation technique of Nei et. al. (1975). Initial effective population size =

4*106, mutation rate = 108, number of iterations = 5000 ..................................... 48
Figure 4. Expected Nei’s D for a population founded at the critical population

density from a founding population at equilibrium (4nv = 0.16) as

determined by habitat size and duration of the bottleneck (A) minimum

density = 0.5/m3; (B) minimum density = 5/m3. Bottleneck duration is the

number of generations population is at minimum effective population

size ........................................................................................................................ 50
Figure 5. Expected F ST for a group of identical populations founded at the critical

density from a single source population as determined by habitat size for

minimum density = 0.5/m3 and 5/m3. See Wade and McCauley (1988) ............. 51

CHAPTER 3: THE EFFECT OF MATING BEHAVIOR AND TEMPERATURE
VARIATION ON THE CRITICAL POPULATION DENSITY OF A FRESHWATER
COPEPOD

Figure 1. Average pre-pursuit swimming velocity (cm/s) for female and male H.
Shoshone. Sample size is indicated inside each bar and error bars represent
the standard deviation. According to ANOVA results, treatment
temperature had a significant effect on both female (p < 0.001) and male
(p<0.001) velocity ................................................................................................. 71
Figure 2. Critical density calculated using male and female velocity from each
temperature treatment and low (1.00065 day'l) or high (1.0043 day") net
reproductive rate. Breeding season length is 60 d ............................................... 72
Figure 3. Sensitivity of estimated critical density to proportional changes in: a)
annual population growth rate, b) swimming velocity, and c) trail length.
Default value is velocity and trail length for males and females at 12°C (see
text), 60 day breeding season, and net reproductive rate = 1.00065 d'l ............... 74
Figure 4. Swimming speed as a function of body size for H. Shoshone (open
shapes) and six other calanoid copepod species (filled circles). Measured
swimming speeds are shown for Sierra Nevada H. Shoshone at 12°C (O),
and Utah H. Shoshone at 12 (c1) and 16°C (A). Data from Tsuda and Miller
(2004), Doall et a1. (1998), Nihongi et a1. (2004), Kierboe and Bageien

(2005), and Sehn et al. (in prep) ........................................................................... 78
APPENDIX A
Figure 1. Average and standard deviation of body length (pm) for adult H.

Shoshone in 2003 (n=14 lakes) and all lake*year combinations (n=40) ............... 89

INTRODUCTION

The ability of a population to recover or re-establish following severe reduction in
size has important influences on species persistence, resistance to disturbance, and, ofien,
colonization success. Inverse density dependence, or the Allee effect, has been suggested
as an important force in the dynamics of low-density populations of an increasing number
of species. Following a severe reduction in population size, species subject to the Allee
effect experience increasing population growth rate with increasing density. This means
smaller populations are less likely to recover, and the Allee effect can produce a
minimum (critical) density below which population growth and persistence is extremely
unlikely. A group of zooplankton, calanoid copepods, are a taxa where the Allee effect is
predicted to occur, based on the mechanism of mate limitation. Because calanoid
copepods occur in three-dimensional habitats, i.e. ponds, lakes and oceans, with
dimensions much greater than their detection and movement capabilities, and reproduce
sexually, individuals in a low-density population may not encounter each other at a high
enough rate to result in population growth. This constitutes mate limitation and creates a
critical density below which the population will decline to extinction.

The theoretical prediction of the Allee effect in calanoid copepods is supported by
observations in nature. The copepod Hesperodiaptomus Shoshone has been extirpated
from shallow alpine lakes following the introduction of non-native fish. Surveys and fish
removal experiments show that following fish removal, this copepod often fails to re-
establish despite recovery of other zooplankton and potential for hatching from a long-
lived egg bank. We hypothesized that this failure to recover is a result of small numbers

of individuals hatching from the egg bank, resulting in a population density below a

critical density established by the Allee effect. Chapter 1 of this dissertation uses
pOpulation-level experiments to detect for the existence of a critical density due to the
Allee effect in the copepod H. Shoshone and calanoid copepods in general. Chapter 2
combines population genetic theory and the Allee effect to establish important
consequences of critical density on population genetics, using H. Shoshone as an concrete
example. Chapter 3 uses individual-scale experiments to predict the critical density based
on physiological and behavioral characteristics of H. Shoshone.

Chapter 1 presents the results of two experiments designed to determine whether
the Allee effect due to mate limitation prevents H. Shoshone populations from recovering
following the removal of fish from alpine lakes. The primary experiment is a whole-lake
reintroduction experiment in six Sierra Nevada lakes to test for a critical density for
population recovery. Stocked copepod densities bracketed a hypothesized critical density
(0.5-5/m3) based on a mechanistic encounter model. The results are consistent with an
Allee effect, with recovery in only 1 lake stocked at high density and population decline
in other lakes. Data on mating probability across experimental and natural H. Shoshone
densities (n=18) shows significant positive correlation with population density,
supporting mate limitation as the mechanism leading to population recovery or decline.
This is one of only a handful of experiments to detect a critical density and establish a
mechanism for the Allee effect and is the only experiment we are aware of that does this
at the scale of natural populations. A second mesocosm experiment conducted in
Michigan with local copepods provided higher replication (n=3 8) and showed increased

growth rate with increasing density, i.e., inverse density dependence. This mesocosm

experiment provides further support for the Allee effect and extends it from single
species to a phenomenon common to calanoid copepods.

Chapter 2 explores the consequences that a critical density has for the population
genetics of calanoid copepods and other species subject to the Allee effect in a similar
way. By setting a lower limit on population size, Allee effects constrain the loss of
genetic variability due to genetic drift. In fact, the minimum population size required by
Allee effects can be much larger than the extremely small population sizes generally
assumed possible in the literature on bottlenecks. As a result, much of the original genetic
variation may be retained in recovered or newly-founded populations. Combining our
findings regarding H. shoshone critical density with reproductive characteristics and
habitat sizes we estimated minimum effective population sizes for H. Shoshone. The
distribution of habitat sizes for this species in the Sierra Nevada suggests that >90% of
populations are resistant to significant changes in heterozygosity or genetic distance and
70-75% of populations will lose <10% of allelic richness during bottlenecks or founder
events. To our knowledge, this is the first consideration of the role of Allee effects in the
loss of genetic variability during population bottlenecks. The results reinforce the fact
that the existence of a critical density can lead to significantly different genetic outcomes
than those predicted when ignoring limits on minimum population size, and suggest Allee
effects could potentially decrease the influence of genetic drift in other taxa.

Chapter 3 presents data from laboratory experiments conducted to improve the
understanding of mate encounter, and therefore the critical density determined by mate
limitation, in H. Shoshone and freshwater calanoid copepods. Advanced imaging

techniques allow characterization and quantification of individual mating encounters.

Earlier work with our collaborators confirmed the ability of male H. Shoshone to increase
mate encounter rate by following pheromone trails produced by females, and we assess
the effect of trail following and swimming speed on critical density. At the same time we
explore the impact of environmental variation, Specifically temperature, on critical
density. Because temperature should impact several parameters of the encounter, such as
swimming speed and pheromone production, trials were conducted over a range of
temperatures applicable to natural environmental variation. These results indicate a large
effect of temperature on swimming speed, and that critical density is more sensitive to
this change in encounter rate than to proportionally larger changes in reproductive rate.
The relationship between temperature, swimming speed and encounter rate suggests an
unconsidered pathway through which temperature changes could affect reproductive rate

in copepods, at least in Species that occur at low densities.

CHAPTER 1
ALLEE EFFECT LIMITS COLONIZATION SUCCESS OF SEXUALLY
REPRODUCING ZOOPLANKTON
ABSTRACT
Understanding the dynamics of populations at low density and Allee effects is a priority
largely due to concern about the decline of rare species and interest in
colonization/invasion dynamics. Despite well-developed theory and observational
support, experimental examinations of the Allee effect in natural systems are rare, partly
because of logistical difficulties associated with experiments at low population density.
We took advantage of fish introduction and removal in alpine lakes to experimentally test
for the Allee effect at the whole-ecosystem scale. The large copepod, Hesperodiaptomus
Shoshone is often extirpated by fish and sometimes fails to recover following fish
disappearance, despite the presence of a long-lived egg bank. Population grth rate of
this dioecious species may be limited by mating encounter rate, such that below some
critical density a colonizing population will fail to establish. We conducted a multi-lake
experiment in which H. Shoshone was stocked at initial densities that bracketed a
hypothesized critical density (0.5-5m'3). Successful recovery by the copepod was
observed in only the lake that was stocked at the highest density. Copepods stocked into
small cages at high density survived and reproduced at rates comparable to natural
populations, confirming that the lakes were suitable habitat for this species. In support of
mate limitation as the mechanism underlying recovery failure, we found a significant
positive relationship between mating success and density across experimental and natural

H. Shoshone populations. Further evidence for Allee effects was obtained from a

mesocosm experiment with a related species, Skistodiaptomus pallidus. Together, these

lines of evidence support the importance of the Allee effect to population recovery of H.
Shoshone in the Sierra Nevada, and to diaptomid copepods in general. This appears to be

the first experimental demonstration of the Allee effect in aquatic animals in nature.

INTRODUCTION

Understanding the dynamics of populations at low density has become a priority,
largely due to concern about rare and endangered species, but also because of the
importance of such dynamics to more general population processes such as colonization.
As a result, inverse density dependence, commonly referred to as the Allee effect (Allee
et a1. 1949, Courchamp et a1. 1999), is being increasingly recognized as an important
phenomenon in natural systems (Stephens and Sutherland 1999, Courchamp et a1. 1999).
Several mechanisms can result in the Allee effect, including mate limitation and obligate
cooperation (Allee et a1. 1949, Odum 1959, Courchamp et a1. 1999), with two related
consequences. The first is a decrease in population growth rate as population density
declines to very low densities. When severe, this decrease in growth rate can result in the
second consequence, a minimum (critical) density below which the population declines to
extinction (Courchamp et a1. 1999). Observational studies have offered evidence for both
decreased growth rate and the existence of critical densities in natural populations (Veit
and Lewis 1996, Kuussaari 1998, Morris 2002, Serrano et a1. 2005, Stoner and Culp
2005), and theoretical models have explored the impact that the Allee effect can have on
population persistence (Boukal and Berec 2002, Dennis 2002, Liebhold and Bascompte
2003, Calabrese and F agan 2004), metapopulation dynamics (Amarasekare 1998, Brassil
2001, Martcheva and Bolker 2007), species invasions (Cruikshank 1999, Drake and
Lodge 2006, Taylor and Hastings 2005), and predator-prey dynamics (Kent et a1. 2003).

Despite long-standing interest in Allee effects, manipulative experiments in
natural systems are rare. A handful of experiments have examined the Allee effect in

plants, finding that low density can threaten population persistence due to insufficient

pollination (Lamont et a1. 1993, Hackney and McGraw 2001) or reduced competitive
ability (Cappuccino 2004). We are aware of only two field experiments on animals that
have tested the existence of a critical density for population establishment, both in insect
populations (Campbell 1976, Berggren 2001). Even in the lab, Allee effects in animals
have only been demonstrated experimentally a few times (Park 1933, Sakuratani et a1.
2001, Noel et a1. 2006).

A unique opportunity to examine Allee effects experimentally in a natural system
arose in the context of fish introduction and removal in alpine lakes of the Sierra Nevada.
Introduction of exotic fish to alpine lakes has led to local extinction of large-bodied
zooplankton, and in some cases cessation of fish stocking or active fish removal has
returned lakes to a fishless state (Knapp et al. 2001b). Following the removal of fish, the
diaptomid copepod Hesperodiaptomus Shoshone sometimes fails to recover despite the
recovery of another extirpated zooplankton species, Daphnia melam'ca (Samelle and
Knapp 2004). Both Daphm'a and copepods have long-lived eggbanks (Hairston and De
Stasio 1988, Parker et a1. 2001, Samelle and Knapp 2004) from which colonists can hatch
alter the lake reverts to a fishless state, but they differ markedly in their mode of
reproduction. Unlike asexually-reproducing Daphm'a, diaptomid copepods are obligately
sexual and do not store sperm (Watras 1983). Moreover, as zooplankton, they inhabit a
three-dimensional bounded habitat that is very large relative to their body size and
mobility. It follows that if densities are too low, the probability of male-female encounter
may be too low to sustain the population, in other words they would be subject to the

Allee effect via mate limitation. As a result, H. Shoshone would need a much larger

number of colonists in a given year in order to successfully re-establish after extirpation
from the water column, relative to Daplmia (Samelle and Knapp 2004).

Because some populations have failed to recover, alpine lakes in the Sierra
Nevada can be used to test the mate limitation hypothesis at the ecosystem scale, via a
whole-lake reintroduction experiment. This study uses a four-tiered approach to examine
the Allee effect in diaptomid copepods. First, we designed a whole lake reintroduction
experiment using multiple lakes from which H. Shoshone had been extirpated, allowing
us to test the effect of initial population density on population recovery across a gradient
of densities. Second, we assessed the evidence for mate limitation and its relationship to
density across our experimental lakes and a set of un-manipulated lakes with established
H. Shoshone populations. Third, we used an in situ caging experiment to determine
whether the experimental lakes are suitable habitat for H. Shoshone reproduction when
mate limitation is alleviated. Fourth, we conducted a mesocosm manipulation of initial
density using a related species of copepod, to allow for more replication and stronger
statistical inference than is possible in whole-lake experiments. The mesocosm
experiment also enabled us to better quantify the relationship between population growth
rate and density in these animals.

METHODS

Hesperodiaptomus Shoshone, is a large (>2mm) deeply-pigmented diaptomid
copepod that occurs in alpine lakes in the Sierra Nevada and the Rocky Mountains from
Colorado to British Columbia. As such, H. Shoshone is highly sensitive to local
extinction after fish introduction, and is largely restricted to fishless lakes (Knapp et a1.

2001). Using paleolimnological techniques (Knapp et al. 2001a) we identified six lakes

in two basins (Fig. 1) from which H. Shoshone had been extirpated by fish (Knapp et al.
2001a, Knapp unpublished data). The lakes had been fishless for a variable amount of
time, ranging from 1 to 25 years (Table 1). Intensive sampling using vertical tows of a 1
m diameter zooplankton net in the weeks prior to the reintroduction, as well as the
previous summer, detected no H. Shoshone in any of the lakes (detection limit = 0.009-
0.015 m'3). Sediment samples detected the presence of viable H. Shoshone eggs within 5
cm of the surface in only 1 of the 6 lakes, No Good Lake (Fig. 1, A) (Samelle and Knapp
2004, Knapp unpublished data). The lakes upstream of Cony and No Good Lakes (Fig.
l, A) both contain established populations of adult H. Shoshone.
Whole-lake experiment

We stocked lakes with H. S/IOS/lone across a density gradient (Table 2) that
bracketed our estimate of the critical density for population establishment (0.5-5 m'3)
(Samelle and Knapp 2004). Estimates of critical density were calculated using
Gerritsen’s (1980) encounter model and estimates of H. Shoshone population growth rate,
length of reproductive season, encounter radius and swimming speed. Our estimates of
critical density are roughly consistent with the range of natural densities of established
populations in Sierra Nevada lakes (6 - 22,000 m’3, Samelle and Knapp 2004). One
would not expect to encounter populations below the critical density, which we have not.
Given the logistic hurdles associated with stocking whole lakes with hundreds of
thousands of copepods, imprecision in the actual number stocked was expected, so we
chose widely spaced target densities, with replicate treatments at the intermediate

densities closest to estimated critical densities (Table 2).

10

 

2kilometers . _ \

 

 

 

 

 

 

..

 

2 kilometers

 

 

 

Figure. 1. Map of experimental lakes. A) Humphreys Basin is on the west side of the
Sierra Crest in the Sierra National Forest, a) Dissertation Lake, source of H. Shoshone
used in reintroduction, 1) No Good Lake, 2) Cony Lake, 3) Square Lake, 4) Knob Lake.
B) Morgan Creek Basin is located on the east side of the crest in the Inyo National Forest
is 12 km north of Humphreys Basin, b) Spire Lake, source of H. Shoshone used in

reintroduction, 5) Finch Lake, 6) Middle Morgan Lake.

11

Amie _. were. BoHESBQQ Ea mm: Emacs woe :6 o 931333: Ba M 8:80 Exam.

 

 

mean—om >8» Zn»: new? 3% Q89

berm 2:250 93 3.3 <oEBm ABJ 35 C3 338% £38
mass?» 33 En: wowmw :w Wm 33
20 Doom 33 #3 woowo Lb m meow
55d 3mm 93 $30 To Mm N2:
003 :8. :6 3R5 To. mu meow
3:0: PM: #3 N33 No V; SS
3. 305mm: 32 0.3 330 w; mb 3.3
Damn—.830: was _.om 2U 20 a 29.9. mm:
meta wmww ado ZU 25 we Em: Eamon"

 

12

Diego N. mxnmaaoaa Qmmwms. amfimzwsm Ea Same“ mflooEDm @3me men one: oxen—.maaaa ES. 9m 3532 cm oouauoam 383 8

was? :5" mama? Em 858 c». Ea man—Sq 8:33? 85 En SEER: moan—m om :8 £855 952.

 

 

flame” 93me moEnm o». man—8a H.838: 9.830:
were Q53 A om noeouoam 3:3»? 25va Oman Ron—Sq
magma Lo wooboo Ummmonmmo: o a: £8
20 Qooa _ w L .80 memonmmo: N E L .28
586 L ooboo Ummmanwmo: m E _ £8
003 o._ :woo Ummmonmao: Pm i _ £8
350: o; wooo mean Nu ENQS

a. 33mm: 0.3 50 were w Ewfiow

 

l3

Late stage H. Shoshone copepodites (pre-adults) to be reintroduced were collected
from two lakes close to the two groups of experimental lakes (Figure 1, Table 1).
“Dissertation” Lake, formally Lake #52121, is fishless, while in Spire Lake, H. Shoshone
co—exist with fish via a deep-water refuge. H. Shoshone are univoltine and emerge from
the egg bank synchronously so we were able to collect large numbers of animals at a
similar life-history stage (i. e, pre-adults) in a short period of time. Animals were
collected using vertical tows of a l m diameter zooplankton net with 350 um mesh. The
collected animals were transferred to a bucket that was mixed, subsampled to estimate
density, and distributed to a set of plastic containers at a density of ~2000 individuals L'l.
The containers were kept cold by covering them with snow, and then transported via a
combination of helicopter and backpack. Time between collection and release of animals
ranged from 2-6 hours and was unrelated to treatment (Table 2). The longest transport
time was for Square Lake, which was stocked at the highest density. Upon delivery,
animals were released throughout the central/deepest part of the lake by emptying the
transport container just below the water’s surface while slowing paddling in a float tube.
Animals were observed swimming in the containers prior to release and in the water
following release.

The low densities associated with tests of the Allee effect necessitated a special
sampling regime with a very low detection limit. Reintroduction lakes were intensively
sampled using vertical tows with a 1m diameter, 350 um mesh zooplankton net.
Detection limits, calculated from the total volume towed during 2-6 visits per lake each
summer were always S 0.005 m}, and often much lower (Table 3). Effort was generally

focused on the deepest area of the lake in order to maximize volume sampled, minimize

l4

Table 3. The volume sampled in each lake each year. The total number of copepods

collected in that volume is in parentheses.

 

Volume sampled each year in m3

(# H. Shoshone collected)

 

Lake 2003 2004 2005 2006
Square 139 (358) 130 (456) 90 (221) 43 (275)
No Good 550 (79) 882 (81) 593 (3) 446 (2)
Knob 766 (49) 413 (0) 363 (0) 427 (0)
Cony 580 (15) 311 (0) 284(2) 280(3)
Finch 368 (5) 300(0) 317(0) 232(0)
M. Morgan Unsampled 198 (0) 214 (0) 211 (0)

 

15

collection of sediment and enable detection of H. Shoshone if they are undergoing vertical
migration. On at least one occasion in each lake horizontal tows and tows in shallow
areas were conducted to make sure animals were not residing in under-sampled areas. H.
Shoshone were never observed in these samples. Density was estimated as the total
number of individuals collected in a lake in a given year divided by the total volume
sampled in that year because individual samples often contained only a handful of
animals.
Assessment of mate limitation

H. Shoshone collected from the experimental lakes, in situ cages (see below)
within the experimental lakes, and eight un-manipulated lakes were examined in order to
assess the relationship between the relative probability of mating and population density.
These samples were collected between 2003 and 2005. Because H. Shoshone are large
and pigmented, the female reproductive state is easily determined. When female
diaptomid copepods are ready to mate (“gravid”), their ripened ovaries become dark and
clearly visible. After mating the fertilized eggs are extruded into an egg sac that is
carried by the female for several days. Females may return to the gravid state before
dropping their eggs sac. Comparing the number of egg—carrying females to the number
of gravid females provides a relative index of mate limitation among populations
(Williamson and Butler 1987).

To maximize confidence in our estimates of mate limitation, each sample was
examined live with the naked eye at the time of collection, preserved, and then re-
examined under the microscope. In both examinations, individuals were scored as

copepodite, male, non-gravid female, gravid female, non-gravid female with eggs, or

16

gravid female with eggs. Live examination was necessary because a portion of females
drop their eggs during preservation and transport, making it impossible to determine
whether unattached eggs observed in the preserved sample were carried by a gravid or
non-gravid individual. Examination under the microscope allowed maturity of
individuals to be confirmed, making certain that animals identified as male, non-gravid
female, and gravid female were correctly characterized. Live and preserved counts were
compared, with close correspondence between the two (less than 2.5% of individuals
mischaracterized).

Reproductive females were scored as mated if egg-bearing, or unmated if gravid
and not bearing eggs. The probability Of mating success of a reproductive female was
modeled as a logistic function of population density. Because conditions such as food
and temperature may influence both proportion of females ready to mate (Williamson and
Butler 1987) and the probability of mating successfully (Chapter 3), and these conditions
will vary among lakes and between years, a random effect of lake by year was included
in the model. Each data point represents one lake-year. For lakes sampled multiple
times in a single year (i. e., experimental lakes), all data were pooled to give a single
estimate of the proportion of females mated across the summer. Density was defined as
adult density since encounters involving immature copepods cannot result in mating
events.

Caging experiment
In 2004, immature copepods were collected from Dissertation Lake and placed in
4, 33 L mesh cages (400 um mesh cylinder 25 cm in diameter, 50 cm long, with 40 cm

conical collector), suspended in each of Dissertation Lake, Knob Lake, and Cony Lake,

17

the latter two being lakes in which copepods Showed no obvious signs of recovery after
the reintroduction in 2003. Animals were subjected to collection and transport conditions
closely matched to the whole-lake experiment, (e. g., confinement at ~2000 L'I for 4
hours, etc.). After animals were added, cages were submerged at 1.5 m below the
surface. Initial density in each cage was 3,333 m'3. After two weeks, cages were
retrieved and all animals and eggs enumerated live. Animals were then returned to the
cages for two additional weeks, after which the contents of each cage was preserved for
microscopic analysis.

Mesocosm experiment

A grid of 38, 1100 L cattle tanks was established at Michigan State University’s
Kellogg Biological Station, Hickory Comers, Michigan in July, 2006. The tanks were
scrubbed with hydrocholoric acid solution and rinsed before use. A thin layer of sand
was placed in the bottom of each tank, and tanks were filled with well water to a volume
of ~1000 L. Tanks were allowed to equilibrate for 4 days, after which nutrients were
added to establish an initially eutrophic environment with a phosphorus concentration of
100 pg L" and an N:P ratio of 30:1 (by atoms) (Hall et a1. 2004). On the same day, an
inoculum of filtered lake water from nearby eutrophic Wintergreen Lake (Mittelbach et
a1. 2006) was added to establish an algal community.

One week after nutrients and algae were added, the small (1.0 - 1.2 mm),
diaptomid copepod, Skistodiaptomus pallidus, was stocked into the tanks. S. pallidus is
common in eutrophic lakes and ponds of the Midwest (Torke 2001). We collected S.
pallidus from Wintergreen Lake with vertical tows of a 30 cm zooplankton net. Animals

were initially isolated from the whole sample using a pipette and then separated by sex

18

into two aquaria. Animals were kept in the aquaria overnight. They were then
transferred into a bucket and concentrated into a flask. Animals were examined
independently by two observers at 10X magnification to confirm sex and eliminate
females that were carrying eggs. Animals were then introduced into the experimental
tanks, with males and females released on opposite sides of the tank. Tanks were
covered with screening to reduce transfer of material between tanks.

Tanks were stocked at the following densities: 2 (n=10), 4 (n=10), 8 (n=8), l6
(n=5), and 32 (n=5) individuals per tank. Replication was greater at low stocking
numbers because we expected greater variability in population growth rate in the lower
density tanks. Sex ratio was always 1:1. The treatments were randomly distributed
across the grid. After 4 weeks, long enough for at least one additional generation to
mature and reproduce (Geiling and Campbell 1971), each tank was sampled by pulling a
30 cm zooplankton net vertically through the tank 3 times, along a transect from one edge
of the tank to the center of the tank.

The entire sample was examined for the presence of S. pallidus and all nauplii,
copepodites and adults were counted. One tank was contaminated with cyclopoid
copepods, but these nauplii and copepodites were easily distinguished from S. pallidus.
Density was estimated as the total of all life stages. Population growth rate (d'l) was then
estimated as:

In Densrtj’fim, — 1n Densztv

. im'n'a/

 

days
After four weeks we failed to detect S. pallidus in 4 tanks: three stocked with 4

individuals, and one stocked with 8. Three of the tanks were immediate neighbors, so it

seems possible that failure to establish could have been a function of factors unrelated to

19

stocked density (all populations established successfully at lower and higher densities).
Statistically we handled these non-detections by adding a constant to the observed density
of zero. The constant added was 1/6 * the detection limit of 3/m3, as recommend by
Mosteller and Tukey (1977). We then calculated a regression analysis of daily growth
rate vs. initial (stocked) density. The results are relatively insensitive to the added
constant, in fact the common practice of adding 1 to calculate the natural logarithms
would have improved the fit in this case.
RESULTS

We successfully introduced live H. Shoshone into the experimental lakes (Figure
2). In the two weeks following stocking in 2003, live animals were detected in all but
Middle Morgan Lake, the lake that was stocked at the lowest density (Table 2). Because
of the amount of time and effort necessary to intensively sample wilderness lakes, we
focused our sampling effort in 2003 on the 5 lakes where we expected H. Shoshone to be
detectable given the limitations on sampling effort (Table 3). Observed density was
significantly less than density expected based on lake volume and estimated inoculum
size (Figure 2). This discrepancy can be attributed to three main causes. Firstly, there
was likely some mortality due to the stress of handling, transport and adjustment to the
new lake (see caging experiment results below). Secondly, natural mortality reduces
population size over the course of the summer even in unmanipulated lakes since all
reproduction results in resting eggs that do not hatch until later years (Appendix A).
Finally, our sampling methods may underestimate density due to net efficiency and
perhaps clearing of the water column during the net lowering phase. However, our

sampling likely reduced population size by no more than 1% on each visit.

20

    
   

 

10.0 I I I I

Target density
l Observed density

E 1.00
it
2:
a)
8 0.10
o

0.01

6
«S

Lake

Figure. 2. Target stocking density and observed density in the weeks following (2003) in

the experimental lakes (logm scale).

21

Population recovery occurred only in Square Lake, the lake with the highest
stocking density (Fig 3). This population remained steady for two years before
increasing from 2005- 2006 by a factor of 2.5 to 6.4 m3, which is within the range
observed for unmanipulated lakes (Samelle and Knapp 2004). The remainder of the
populations declined after stocking (Fig. 3). The No Good Lake population, stocked at 1
rn'3 but with an observed first-year density of 0.14 m'3, exhibited an exponential decline
from the initial stocked population size (Fig 3). H. Shoshone were detected in 2005 and
2006 in Cony Lake, which was stocked at 0.1 rn'3 (Fig 3), but these rare individuals
(Table 3) probably do not constitute a reproducing population, as discussed further
below.

The degree of mate limitation was positively dependent on population density
across stocked and established populations (Fig. 4). Mate limitation was also highly
variable in established populations such that the lowest values of mating success in
established populations were similar to the highest values in the stocked populations,
despite the fact that population densities in stocked lakes were more than an order of
magnitude lower.

The results of the caging experiment (Table 4) showed that H. Shoshone were able
to reproduce successfully when caged at high density (3333 m3) in Knob lake and Cony
Lake, at the same time that stocked populations in the lakes declined (Fig. 3, Table 4).
Clutch size and clutches per female for the caged copepods in Knob Lake and No Good
Lake were similar or higher than those observed in caged and uncaged copepods in the

source lake (Dissertation Lake) and in Square Lake (Table 4). Mortality was also similar

22

 

 

 

 

 

10-0 : I I I I I I 5
6‘ P ‘
g - .
$51.00 :- -,
3‘ E 2005:00051 5
'IT) - 2006 = 0.0045 -
g . - Year
0 i 2004
8 0-10 5' — - - 2005
E -’- i '5 2006
a) u- : ..
<0 r , -
'0 I- '7‘. q
0
0.01 5- l '5
3 I I I I l 1

 

 

 

 

 

>0<.\O1 0.02 0.03 0.07 0.15 2.6 \
e5 9 or ti nli it=0.005
s- 9°

Initial Density (#lm3)

Figure. 3. Observed densities in the experimental lakes (logm scale) immediately
following reintroduction (2003) and in the following years. Lakes are identified by target
density and are in the same order as Fig. 2. Detection limit < 0.005 m'3 unless noted

otherwise.

23

 

 

 

 

 

I 1IIIIIII I I IIIIIII I I IIIIIII I I IIIIIII I I IIIIIII I I I IIIIII
1'0 —0 Natural population 00 _
A Reintroduced population
9 0.8 -
(0
E
Q 0.6 — —
U)
(D
.c:
.9 0.4 " -
2
O
0.2 — “
0.0A 1 1A1] l llllllll I llllllll l llllllll l [4111111 I lllllll

10'2 10'1 100 101 102 103 104

Adult Density (#lm3)

Figure. 4. Proportion of reproductive females bearing eggs vs. density of adult copepods
in reintroduced populations (A) and natural populations (0). Filled triangles ( A)
represent Square Lake, the reintroduction lake where the population recovered. The line
is the logistic regression function: WithEggs = (eA-l.184 + 0.74 * AdultDensity) / (1 +

(eA-1.184 + 0.74 * AdultDensity)) (p < 0.0005).

24

Table 4. Reproductive and survival data from the caging experiment, the source of the

animals used in the caging experiment (Dissertation Lake), and the successfully

reintroduced population (Square Lake). Clutch size is average and standard deviation of

eggs per egg bundle over the course of the experiment (cages) or the summer of 2004

(Dissertation and Square Lake). # clutches/ female is the number of clutches per

reproductive (gravid and/or egg-bearing) female for 2 cages per lake after 2 weeks in the

cages (cages) or for the single closest sampling date (lakes). All sampling took place

between 7/19/2004-7/21/2004. Mortality is the average daily mortality calculated from

the regression of ln(density) vs. time. A-Morta1ity is mean of 4 cages sampled on 3 dates

from 7/11/2004-8/8/2004. B- Dissertation Lake mortality is from 3 sampling dates from

7/29/2004 — 8/22/2004, equivalent data is not available for same time period as the cages.

C — Square Lake mortality data is from 4 sample dates.

 

 

Location Mean clutch #clutches/ female Daily mortality
size (s.d) (s.d.) (s.d)
Knob (cages) 9.86 (0.75) 0.39 (0.22) 0.040 (0.004)A
Cony (cages) 15.25 (0.74) 0.20 (0.03) 0.033 (0.007)A
Dissertation (cages) 13.50 (0.51) 0.44 (0.08) 0.013 (0.004)A
Dissertation (lake) 12.38 (0.06) 0.27 (NA) 0.074B
Square (lake) 19.49 (1.29) 0.22 (NA) 0.112C

 

25

to the source lake and substantially lower than the mortality observed following the 2003
reintroduction in Square Lake (Table 4).

Results from the mesocosm experiment provide clear evidence of inverse density
dependence (i.e., the Allee effect) in the Midwestern diaptomid, S. pallidus (Fig. 5). In
contrast to the whole-lake experiment, S. pallidus successfully colonized most of the
mesocosms, including all at the lowest stocking density of 2 m'3, but population growth
rate increased significantly with stocking density (p=0.025, R2=0.13). Mean growth rate
at the highest stocking density was 0.214, which is below the reported range of 025-04
for maximum copepod growth rates (Allan 1976). The observed differences in growth
rate resulted in two orders of magnitude difference in mean population size over the 1
month-long experiment, with mean population size = 114 / m3 ($49) when 2 individuals
were stocked and mean population size = 12200 / m3 ((i6600) when 32 individuals were
stocked.

DISCUSSION
This paper presents four lines of evidence that the Allee effect operates to reduce re-
establishment success after local extinction in diaptomid copepods. In the whole-lake
experiment, we found recovery only in Square Lake, the lake with the highest density of
reintroduced animals (Fig 3). Copepods caged at high density in two of the lakes with
failed recovery survived and reproduced at rates comparable to or higher than in natural
populations (Table 4). This suggests that recovery failure was not due to the
experimental lakes being unsuitable habitat for H. Shoshone, the most likely alternative

hypothesi

26

 

0-3 I F I

Growth rate (daily)

 

 

 

'01 (LI I I

0 V10 20 30 40
Stocked Density

 

Figure. 5. Daily population growth rate vs. introduced density of S. pallidus in 1000 L
mesocosms. The line is the linear regression function: y = 0.003*x + 0.092, p = 0.025, R2
= 0.13, n = 38. Final density was undetectable in four cases resulting in negative growth

rate (see Methods). Experiment lasted for 4 weeks.

27

s for recovery failure. We expected all the lakes to be suitable habitat for H. Shoshone
because the species was present in the past (Knapp et al. 2001a, Knapp unpublished
data), but it is possible that the lakes could have become unsuitable over the last several
decades as a result of climate change or some other factor. Our caging results, and the
success of H. Shoshone in Square Lake, provide evidence against this hypothesis.
Further, pelagic predators such as cyclopoid copepods and Chaoborus (Parker et a1.
2001) are absent from these lakes (Samelle and Knapp 2004), so caged animals would
not be expected to have higher survival relative to free-swimming animals. In addition,
the degree of mate limitation was greatly intensified at the low population densities of the
experimental lakes (Fig 4). Female mating success was comparable to levels found in
natural populations only in Square Lake, the only experimental lake showing evidence of
successful recovery. Taken together, these three lines of evidence strongly suggest that
the primary reason for recovery failure of H. Shoshone in 5 of the 6 experimental lakes
was reduction in population growth rate below zero by insufficient mating success.
Finally, the results of the mesocosm experiment with S. pallidus were congruent with the
alpine lake data in that population growth rate of a diaptomid copepod was found to be a
positive function of initial density (Fig. 5).

Our detection of isolated, rare H. Shoshone individuals in Cony and No Good
Lakes (Fig. 3, Table 3) does not seem to represent ongoing recovery of these two
populations. No mated females have ever been detected in Cony Lake, and mated
females were not detected in 2005 or 2006 in No Good Lake (Fig. 4). The individuals
detected by our sampling could instead arise from viable resting eggs that were deposited

prior to fish presence or from downstream transport of eggs or live copepods. No Good

28

Lake has viable eggs in the top 5 cm of sediment (700 m'z, Samelle and Knapp 2004),
and while eggs were not detected in the top 5 cm of Cony Lake sediment, our intensive
sampling disturbs sediment across a wide swath of the deepest part of the lake, and the
disturbance could extend more than 5 cm into sofi sediment at the bottom of these lakes
(Kramer personal observation). Additionally, the upstream lakes contain H. Shoshone
(see Methods), and while downstream transport of adults is unlikely (Kramer,
unpublished data), egg transport may be more common. The fact that H. Shoshone have
not recovered in No Good Lake in the 5 years since fish removal (Fig. 3, Samelle and
Knapp 2004) despite these possible inputs suggests that these isolated detections are not a
prelude to population recovery.

The results of the whole-lake experiment are congruent with studies on H.
arcticus, an ecologically similar and closely-related species in the Canadian Rockies. H.
arcticus is also driven locally extinct by the stocking of non-native fish (Parker 1996),
followed in some cases by recovery failure where the egg bank is depleted (Parker et al.
1996). H. arcticus was successfully re-established in a single lake through the
introduction of reproductive adults at a density of 1.5 m'3 (McNaught et al. 1999), which
is within the range of H. Shoshone critical densities suggested by our experiment (see
below). However, the stocking density used for H. arctz’cus is not necessarily
comparable to our reintroduction densities because H. arcticus females had likely already
mated at the time of collection (McNaught et al. 1999).

Although the evidence supporting operation of the Allee effect via mate limitation
in diaptomid copepods is strong (Figs 3 - 5), estimating the critical density below which

populations will fail to establish is less certain. If we assume that the density we

29

observed over the weeks immediately following stocking (not actual numbers stocked) is
the more relevant measure of initial density, the results of the whole-lake experiment
suggest a critical density between 0.2 111’3 (No Good Lake) and 3 m"3 (Square Lake) for H.
Shoshone (Fig 3), which is slightly lower than we originally estimated based on
laboratory data alone (0.5 - 5 m3). Calculating lake densities based on the entire lake
volume as we have done, rather than for the pelagic zone only, seems justified because
we regularly observed H. Shoshone in the littoral zone in established populations. For S.
pallidus, the tanks were apparently too small to permit an accurate estimate of critical
density, since copepods established in all tanks stocked at 2 m'3 (Fig. 5). Based on
encounter models (Gerritsen 1980, Kiorboe and Bagioen 2005), S. pallidus should have a
higher critical density due to its smaller size, which reduces swimming speed and
detection radius. However, decreased encounter rate could have been offset by increased
population growth rate in the warmer and more productive mesocosms. In addition, S.
pallidus produces eggs that hatch immediately, in contrast to the exclusive production of
resting eggs by H. Shoshone, which also contributes to higher population growth, and
therefore lower critical density. In cases were an egg bank is absent, such as Square
Lake, resting eggs will become buried by bioturbation (Keams et al. 1996) without
concurrent uncovering of eggs already in the egg bank, until the egg bank is re-stocked.
This potentially explains the slow population growth observed in Square Lake (Fig. 3),
despite our estimate that the Square Lake population was able to produce 5.6 eggs
individual'l when mating success was 20%. S. pallidus’ production of immediately
hatching eggs avoids this brake on population growth. This difference in life history

could be important in colonization success between diaptomid species, and could also

30

cause colonization success to vary seasonally in species which switch between the two
egg types (Ellner et al. 1999).

The results in this paper support the hypothesized explanation for the pattern of H.
Shoshone recovery in alpine lakes following fish disappearance (Samelle and Knapp
2004.). H. Shoshone failed to recover in 46 % of 41 Sierra lakes it inhabited before fish
introduction, a rate much lower than that of Daphnia melanica, and failure was correlated
with fish residence time (Knapp and Samelle in prep). During fish presence the egg bank
becomes depeleted over time due to hatching and egg mortality, so longer fish residence
time leads to a lower density when fish disappear. If that initial density is much below ~1
m'3, our whole-lake results suggest that recovery is unlikely.

If H. Shoshone are unlikely to recover from depleted, yet viable, egg banks or
colonization events consisting of thousands of viable individuals, how did they become
so widespread in the first place? In the Sierra Nevada, H. Shoshone inhabit ~60% of
never-stocked and stocked-now-fishless lakes (Knapp et al. 2001b). Our results suggest
that these lakes required many thousands of individuals to be present simultaneously for
recovery to be probable, and that larger lakes would require many more. Resting eggs
are likely the key to resolving this paradox. One possible explanation is that large
numbers of eggs can be transferred by events that are either rare, or the result of past
transport mechanisms that are no longer relevant. Another possibility, that we plan to
explore further, is whether the egg bank provides a way to accumulate the reproductive
effort of repeated, individual colonization events even if no single event exceeds the

critical density. This explanation would require that a substantial fraction of resting eggs

not hatch in the first few years after deposition (to allow a build-up of individuals) and

31

would predict that establishment would take many years or even decades. Alternatively,
it is possible that copepods spread through surface water at a time when the lakes were
more connected, as argued by Stemberger (1995) for lowland lakes. However, this
explanation seems less likely for high—altitude lakes (Stemberger 1995). Finally, perhaps
the widespread distribution of H. slzoshone, is the result of a highly stochastic process
with eventual colonization success following one of many repeated dispersal events, with
each individual event having a very low probability of success because of mate
limitation.

This study highlights the difficulty of testing the Allee effect, and estimating
critical density in particular. At spatial scales that are substantially smaller than whole
lakes, the critical density would be difficult or impossible to assess. This is illustrated by
the results of our mesocosm experiment where the tanks were too small to measure
critical density in a small copepod.

Difficulty of detection does not mean the Allee effect is unimportant. This study
shows that Allee effects are an important force in a range of dynamics, even in species
known for large abundances. Natural populations of diaptomid copepods are in the
millions even in small habitats, yet our study suggests Allee effects can play an important
role in the establishment and persistence of each of those large populations. We believe
this to be the first experimental evidence for the Allee effect in an aquatic animal and it
provides experimental confirmation that theoretical models of the Allee effect have
relevance to natural populations. These results are relevant to other taxa expected to be
subject to mate limitation in particular, including conch (Stoner and Ray-Culp 2000), and

several insects (Hopper and Roush 1993, Berggren 2001, Liebhold and Bascompte 2003)

32

and to the Allee effect in general. These experimental results reinforce the importance of
considering and testing Allee effects in the context of the recovery of populations
following anthropogenic disturbance, natural colonization processes and species
invasions.
ACKNOWLEDGEMENTS

We thank J. Garton, C. Archer, B. Wilson, G. Goldsmith, C. Brownfield, J.
Nicholl and T. Armstrong for assistance with the fieldwork that made this study possible.
Research and collecting permits were provided by the Inyo and Sierra National Forests,
and the California Department of Fish and Game. The work in the Sierra Nevada was
supported by a National Science Foundation Graduate Research Fellowship, National
Science Foundation grants DEB-9629473, DEB-0075509, an REU supplement, VESR
Graduate Student grant, and a Sigma-Xi Grant-in-aid of Research . We thank M. Leibold,
J. Pantel and G. Mittelbach for lending their equipment and expertise to the mesocosm
experiment, S. Smith for locations of diaptomid copepod populations, and the Michigan
State University Ecology, Evolutionary Biology and Behavior program for fimding

support for the mesocosm experiment.

33

LITERATURE CITED

Allan, J. D. 1976. Life-history patterns in zooplankton. American Naturalist 1102165-
180.

Allee, W. C., A. Emerson, et al. 1949. Principles of animal ecology. Philadelphia,
Saunders.

Amarasekare, P. 1998. Allee Effects in Metapopulation Dynamics. The American
Naturalist 152:298-302.

Berggren, A. 2001. Colonization success in Roesel’s bush-cricket Metrioptera roeseli:
the effects of propagule size. Ecology 82:274-280.

Boukal, D. S. and L. Berec. 2002. Single-species models of the Allee effect: Extinction

boundaries, sex ratios and mate encounters. Journal of Theoretical Biology 2182375-
394

Brassil, C. E. 2001. Mean time to extinction of a metapopulation with an Allee effect.
Ecological Modelling 14329-16.

Buskey, E. J ., J. O. Peterson, and J. W. Ambler. 1996. The swarming behavior of the
copepod Dioithona oculata: In situ and laboratory studies. Limnology and
Oceanography. 41: 513—521 .

Calabrese, J. M. and W. F. Fagan. 2004. Lost in time, lonely, and single: reproductive
asynchrony and the Allee effect. American Naturalist 164225-37.

Campbell, M. M. 1976. Colonisation of Aphytis melinus DeBach (Hymenoptera,

Aphelinidae) in Aonidiella aurantii (Mask) (Hemiptera, Coccidae) on citrus in South
Australia. Bulletin of Entomological Research 65:659-668.

Cappuccino, N. 2004. Allee effect in an invasive alien plant, pale swallow-wort
Vincetoxicum rossicum (Asclepiadaceae). Oikos 106:3-8.

Courchamp, F ., T. Clutton-Brock, and B. Grenfell. 1999. Inverse density dependence and
the Allee effect. Trends in Ecology and Evolution 14:405-410.

Cruickshank, I., W. S. C. Gurney, and A. R. Veitch. 1999. The characteristics of
epidemics and invasions with thresholds. Theoretical Population Biology 56:279-292.

Dennis, B. 2002. Allee effects in stochastic populations. Oikos 96:389-401.
Doall, M. H., S. P. Colin SP, J. R. Strickler, and J. Yen. 1998. Locating a mate in 3D:

the case of T emora longicornis. Philosophical Transactions of the Royal Society B:
Biological Sciences 353:681-689.

34

Drake, J. M. and D. M. Lodge. 2006. Allee effects, propagule pressure and the
probability of establishment: Risk analysis for biological invasions. Biological
Invasions 8: 365-375.

Ellner, S. P., N. G. Hairston, C.M. Keams, and D. Babai. 1999. The roles of fluctuating
selection and long-term diapause in microevolution of diapause timing in a freshwater
copepod. Evolution 53:1 11-122.

Geiling, W. T. & R. S. Campbell, 1972. The effects of temperature on the development

rate of the major life stages of Diaptomus pallidus Herrick. Limnology and
Oceanography 17: 304—307.

Gem'tsen, J. 1980. Sex and parthenogenesis in sparse populations. American Naturalist
115:718—742.

Hackney, BE. and J. B. McGraw. 2001. Experimental demonstration of an Allee effect in
American ginseng. Conservation Biology 15:129-136.

Hairston, N. G., and B. T. J. De Stasio. 1988. Rate of evolution slowed by a dormant
propagule pool. Nature 336:239—242.

Hall, S. R., M. A. Leibold, D. A. Lytle, and V. H. Smith. 2004. Stoichiometry and
planktonic grazer composition over gradients of light, nutrients, and predation risk.
Ecology 85:2291-2301.

Hopper, K. R. and R. T. Roush. 1993. Mate finding, dispersal, number released and the
success of biological control introductions. Ecological Entomology 18:321-331.

Keams, C. M., N. G. Hairston, and D. H. Kesler. 1996. Particle transport by benthic
invertebrates: Its role in egg bank dynamics. Hydrobiologia 332:63-70.

Kent, A., C. P. Doncaster, and T. Sluckin. 2003. Consequences for predators of rescue
and Allee effects on prey. Ecological Modelling 162:233-245.

Knapp, R. A., J. A. Garton, and O. Samelle. 2001 a. The use of egg shells to infer the
historical presence of copepods in alpine lakes. Journal of Paleolimnology 25:539-

543.

Knapp, R. A., K. R. Matthews, and O. Samelle. 2001b. Resistance and resilience of
alpine lake fauna to fish introductions. Ecological Monographs 71 :401-421.

Kuussaari, M., I. Saccheri, M. Camara, and I. Hanski. 1998. Allee effect and population
dynamics in the Glanville fritillary butterfly. Oikos 82:384-392.

35

Lamont, B. B., P. G. L. Klinkhamer, and E. T. F. Witkowski. 1993. Population
fragmentation may reduce fertility to zero in Banksia goodii: a demonstration of the
Allee effect. 0ecologia 94:446-450.

Leising, A.W., and J. Yen. 1997. Spacing mechanisms within light-induced copepod
swarms. Marine Ecology-Progress Series 155: 127-135

Liebhold, A. and J. Bascompte. 2003. The Allee effect, stochastic dynamics and the
eradication of alien species. Ecology Letters 6: 133-140.

Martcheva, M. and B. M. Bolker. 2007. The impact of the allee effect in dispersal and
patch-occupancy age on the dynamics of metapopulations. Bulletin of Mathematical
Biology 69:135-156.

McNaught, A. S., D. W. Schindler, B. R. Parker, A. J. Paul, R. S. Anderson, D. B.
Donald, and M. Agbeti. 1999. Restoration of the food web of an alpine lake following
fish stocking. Limnology and Oceanography 44:127-136.

Mittelbach, G. G., E. A. Garcia, and Y. Taniguchi. 2006. Fish reintroductions reveal
smooth transitions between lake community states. Ecology 87 :312-3 1 8.

Morris, D.W. 2002. Measuring the Allee effect: Positive density dependence in small
mammals Ecology 83: 14-20

Mosteller, F. and J. W. Tukey. 1977. Data analysis and regression: a second course in
statistics. Addison-Wesley, Reading, Mass.

Noel, H. L., S. P. Hopkin, T. H. Hutchinson, T. D. Williams, and R. M. Sibly. 2006.
Towards a population ecology of stressed environments: the effects of zinc on the
springtail F olsomia candida. Journal of Applied Ecology 43:325-332.

Odum, P. E. 1959. Fundamentals of Ecology. Saunders, Philadelphia.

Park, T. 1933. Studies in population physiology 11. Factors regulating initial grth of
T ribolium confusum populations. Journal of Experimental Zoology 65: 1 7-42.

Parker, B. R., F. M. Wilhelm, and D. W. Schindler. 1996. Recovery of
Hesperodiaptomus arcticus populations from diapausing eggs following elimination
by stocked salmonids. Canadian Journal of Zoology-Revue Canadienne de Zoologie
74:1292-1297.

Sakuratani, Y., K. Nakao, N. Aoki, and T. Sugimoto. 2001. Effect of population density

of C ylas formicarius (F abricius) (Coleoptera : Brentidae) on the progeny populations.
Applied Entomology and Zoology 36:19-23.

36

Samelle, O., and R. A. Knapp. 2004. Zooplankton recovery after fish removal:
Limitations of the egg bank. Limnology and Oceanography 49: 1382-1392.

Serrano, D., D. Oro, U. Esperanza, and J. L. Tella. 2005. Colony size selection
determines adult survival and dispersal preferences: Allee effects in a colonial bird.
American Naturalist 166:E22-E3 l.

Stemberger, R. S. 1995. Pleistocene refuge areas and postglacial dispersal of copepods of
the northeastern United States. Canadian Journal of Fisheries and Aquatic Sciences
52:2197-2210

Stephens, P. A., and W. J. Sutherland. 1999. Consequences of the Allee effect for
behaviour, ecology and conservation. Trends in Ecology and Evolution 14:401-405.

Stoner, A. W. and M. Ray-Culp. 2000. Evidence for Allee effects in an over-harvested
marine gastropod: density-dependent mating and egg production. Marine Ecology-
Progress Series 202:297-302.

Taylor, C. M. and A. Hastings. 2005. Allee effects in biological invasions. Ecology
Letters 82895-908.

Torke, B. 2001. The distribution of calanoid copepods in the plankton of Wisconsin
Lakes. Hydrobiologia 453:351-365.

Veit, R. R. and M. A. Lewis. 1996. Dispersal, population growth, and the Allee effect:
dynamics of the house finch invasion of eastern North America. The American
Naturalist 148: 255-274

Watras, C. J. 1983. Mate location by diaptomid copepods. Journal of Plankton Research
52417-425.

Williamson, C. E., and N. M. Butler. 1987. Temperature, food and mate limitation of

copepod reproductive rates: separating the effects of multiple hypotheses. Journal of
Plankton Research 9:821 -836.

37

CHAPTER 2
LIMITS TO GENETIC BOTTLENECKS AND FOUNDER EVENTS IMPOSED
BY THE ALLEE EFFECT
ABSTRACT
Allee effects create a minimum (critical) density for population persistence when
population growth rate becomes negative at low population density. Critical density and
habitat size combine to define a minimum population size during population bottlenecks
or founder events. By setting a lower limit on population size, Allee effects constrain the
loss of genetic variability due to genetic drift. As a result, most of the original genetic
variation may be retained in recovered or newly-founded populations. In the alpine
copepod, Hesperodiaptomus shoshone, models, surveys and experimental data suggest
populations at densities less than 0.5 - 5 individuals/m3 are unable to persist due to mate
limitation. Combining this range of critical density with reproductive characteristics and
the distribution of habitat sizes in nature, we estimated minimum effective population
sizes for H. shoshone. Our calculations suggest that >90% of H. shoshone populations in
the Sierra Nevada are resistant to significant changes in heterozygosity or genetic
distance, and 70-75% of populations will lose <10% of allelic richness, during

bottlenecks or founder events. This appears to be the first attempt to account for Allee

effects as a factor limiting the loss of genetic variability during population bottlenecks.

38

INTRODUCTION

The Allee effect, viewed broadly, is a decrease in population growth rate as
population density declines (inverse density-dependence), resulting from mechanisms
such as mate limitation, obligate cooperation, and group-based antipredator strategies
(Allee et al. 1949, Odum 1959, Courchamp et al. 1999). A specific consequence of
strong inverse density dependence is the existence of a critical population density, the
density below which population growth rate is negative (Courchamp et a1. 1999). In
effect, critical density represents a minimum population density, and depending on
habitat size, a minimum population size, because at lower densities both the persistence
of established populations and the establishment of new populations will be unlikely. To
date, most discussions of the Allee effect have focused on issues relating to the
establishment of invasive species (Lewis and Kareiva 1993, Veit and Lewis 1993, Leung
et al. 2004), or to population decline and recovery of endangered species (Kuussaari et al.
1998, Lamont et al. 1993, F orsyth 2003) or collapsed populations of harvested taxa
(Myers et al. 1995, Liennann and Hilbom 1997, Gascoigne and Lipcius 2004). Here, we
draw attention to an unappreciated connection between critical density, a phenomenon
drawn from population dynamics, and loss of genetic diversity in small populations, a
phenomenon drawn from population genetics.

By constraining the minimum effective population size, the Allee effect will also
constrain the genetic outcome of bottleneck events. Here we define a bottleneck event as
a severe decrease in population size or the founding of a new population by a small
number of colonists. Theory predicts that for sufficiently severe bottlenecks, the change

in gene frequency depends only on the effective population size during the bottleneck and

39

the duration of the bottleneck, because genetic drift overwhelms selection in small
populations (Nei et al. 1975, Maruyama and F uerst 1985). Because many populations
rebound rapidly from small size, with threatened and endangered species a notable
exception, the minimum effective population size sustained during a bottleneck event
largely determines the genetic makeup of the post-bottleneck population, affecting both
the absolute genetic variation remaining and the degree of genetic divergence from the
pre-bottleneck population’s genetic state (Nei et al. 1975, Wade and McCauley 1988).
The theoretical and empirical literature tends to focus on extreme bottlenecks with
very small effective population sizes. The minimum effective population size in
theoretical analyses is often ten or fewer individuals (Nei et al. 1975, Pannell and
Charlesworth 1999, Whitlock and McCauley 1990, Leberg 2002). Empirical work,
especially on species of conservation concern, has confirmed both the occurrence and the
genetic impact of such extreme bottlenecks (Le Page et al. 2000, Robichaux et al. 1997,
Glenn et al. 1999). As a result, it is typical for observations of very low genetic
variability in natural populations or high genetic divergence between populations to be
automatically ascribed to genetic drift associated with severe population bottlenecks. In
effect, most analyses of genetic variability and genetic differentiation via drift contain an
implicit assumption that the population could have established or recovered from an
arbitrarily small size; 1 or 2 individuals in the case of asexual or sexual species,
respectively. This assumption affects not only direct assessments of population
bottlenecks, but also analyses which use genetic differentiation to infer other processes
such as migration (Boileau et al. 1992, Ramstad et al. 2004). If the minimum effective

population size determined by the Allee effect is relatively large, as suggested for some

40

species such as zooplankton (Gerritsen 1980), conch (Stoner and Ray-Culp 200) and
several insects (Hopper and Roush 1993, Berggren 2001, Liebhold and Bascompte 2003),
bottlenecks or founder events may predominantly result in little loss of genetic variability
via drift.

Because the Allee effect is a function of population density, habitat size is the key
to predicting minimum possible population size, and therefore the change in genetic
variability and divergence, for species subject to an Allee effect. We present a
quantitative analysis of how habitat size constrains the loss of genetic diversity and the
increase in genetic divergence in a diaptomid copepod, a group of organisms potentially
subject to strong mate limitation. These small planktonic crustaceans reproduce sexually
(unlike cladocerans such as Daphnia) and have limited mobility in a potentially gigantic
habitat of a lake or ocean, These constraints should create a minimum critical density
below which the probability of encountering a mate is too low for a population to persist
(Gerritsen 1980). Data confirms the existence of mate limitation in diaptomid copepods
(Chapter 1), and there is observational and experimental support for the existence of a
critical density (Samelle and Knapp 2004, Parker et al. 1996, McNaught et al. 1999,
Chapter 1). Copepod critical density can be estimated from information on body size,
swimming speed and reproductive characteristics (Gerritsen 1980, Samelle and Knapp
2004). Because copepods live in lakes and ponds, as well as the ocean, we can also
define and quantify habitat sizes encountered in nature. We use this information to assess
the relationship between habitat size and the genetic impact of bottleneck events and
determine if Allee effects should be considered when analyzing population genetic data

of these copepods and other species with analogous characteristics.

41

METHODS

Our analysis focuses on Hesperodiaptomus shoshone, a large (>2mm) diaptomid
copepod that occurs in alpine lakes in the Sierra Nevada and the Rocky Mountains from
Colorado to British Columbia. H. shoshone is univoltine and susceptible to eradication
by fish introduction (Knapp et al. 2001). The critical density of H. shoshone has been
estimated by Samelle and Knapp (2004). Using a range of values for population grth
rate and length of reproductive season and Gerritsen’s (1980) encounter model, they
predicted a critical density of 0.5-5/m3 (Samelle and Knapp 2004). This estimated range
is supported by observations on natural populations (Samelle and Knapp 2004) as well as
data from experimental reintroductions (Chapter 1). Critical density multiplied by habitat
size (in this case lake volume) gives the minimum population size. For example, an
average size H. shoshone lake in the Sierra Nevada has a minimum population size of
64,000-640,000 individuals.

The range of relevant habitat sizes was determined using the surface area and
maximum depth of 169 H. shoshone-containing lakes which constituted a representative
sample of H. shoshone habitat in the Sierra Nevada (Knapp, unpublished data). We
estimated mean depth from maximum depth with the regression equation: mean depth =
0.782 + 0.315 * max depth (R2=0.77), which is based on 2345 northern U.S. lakes
spanning a range of maximum depths similar to the Sierra Nevada lakes (Soranno,
personal communication). Volume was then estimated as surface area * mean depth.

The key link in examining the relationship between minimum population size and
population genetics is the effective population size. Changes in gene frequency due to

drift are determined by effective population size rather than census population size

42

(Wright 1931), and fluctuations in population size, skewed sex ratios and variance in
reproductive success act to reduce the effective population size (Wright 1938). We used
life history data and made several assumptions to develop a simple, deterministic
estimate of the effective population size of H. shoshone populations at the critical
density. First we assume population size during the bottleneck remains at the minimum
and that the sex ratio of H. shoshone is 1:1 (Appendix A). Therefore, any reduction in
effective population size is due to variance in reproductive output.

We expect high variance in reproductive success because the low encounter rate
at critical density will lead to zero reproductive output in the majority of individuals. To
estimate the number of individuals that do contribute offspring to the next generation, we
use the clutch size (mean = 16.02, sd = 4.18) measured for H. shoshone stocked into
mesh enclosures in 4 lakes (Appendix A), and the conservative assumption that
successful individuals (male and female) are able to mate twice and each of these females
produce two clutches during the reproductive season. Our aim is to err towards
overestimating variance in reproductive success by assuming all reproduction is
contributed by a somewhat unrealistically limited pool of individuals, but this may be
somewhat offset by an assumption of zero variance in reproduction among successful
pairs. Based on the assumptions above, we can estimate the minimum number of
reproducing individuals (NC) necessary to maintain population size at the critical density
as follows:

N _ habitat size x critical density _ habitat size x critical density

 

 

clutch size x #clutches/pair 16 x 2

We also conservatively assume 100% hatching success and zero mortality before

reproduction.

43

We consider bottlenecks that last for l to 10 generations. Longer bottlenecks are
unlikely for H. shoshone and diaptomid copepods in general (Samelle and Knapp 2004,
McNaught et al. 1999). McNaught et al. (1999) found population growth to be
exponential, with natural population density nearly attained in only 4 generations
[following reintroduction at a density of 1.5 individuals / m3]. Short bottleneck duration
is also a likely scenario for any species with high reproductive rates not subject to habitat
loss, harvest or other external forces.

The ranges of bottleneck duration and minimum effective population size
determined above are used to examine how two measures of genetic variability and two
measures of genetic divergence are affected by bottlenecks in populations subject to
Allee effects. Change in genetic variability is estimated by heterozygosity; perhaps the
most widely used measure for the maintenance of genetic variation, and allelic richness,
which is much more sensitive to changes in population size (Nei 1975, Nei 1987). For
reasons of simplicity and consistency with existing analyses, we estimate change in
heterozygosity and allelic richness as a proportion of the variation present before the
bottleneck, as in Nei et al. (1975). We calculate both measures in the same way as Nei et
al. (1975), using their analytical solution for heterozygosity and their simulation method
for allelic richness (see Nei et al. 1975).

To represent genetic divergence we consider Nei’s genetic distance (D) because it
allows us to consider two currently or historically linked populations (Chakraborty and
Nei 1977). We estimate D between two populations, where one is founded by a random
sample from a second population which is at mutation-drift equilibrium, using the infinite

alleles model as in equation (21) from Chakraborty and Nei (1977). We also consider

44

F g, which measures the overall genetic similarity of a group of populations, following
the assumptions for the migrant pool model and a small number of colonists as in
equation (1) from Wade and McCauley (1988). Under this model FST measures the
genetic differentiation among a group of populations founded from a single source
population (Wade and McCauley 1988). This model was selected because it is the only
formulation which does not require parameters, such as migration, which are not present
in our other analyses.

RESULTS

The sample of lakes inhabited by H. shoshone in the Sierra Nevada had volumes
ranging from 77 m3 to 7.3 * 10° m3, with a median volume of 1.3 * 105 m3 (Fig. 1). The
corresponding range of minimum effective population size for a minimum population
density of 0.5/m3 is 2.4 — 2.3 * 105 individuals, or 1.2 — 1.15 * 105 pairs, with a median
effective population size of 4.0 * 103 (Fig. 1). With a critical density of 5/m3 the lower
limit on effective population size for similarly-sized habitat is ten times higher.

The proportion of the original heterozygosity retained in a bottlenecked
population rises rapidly as habitat size increases from the minimum habitat size (Fig. 2).
It asymptotes at a small habitat size; over 90% of the original variability is retained in
habitats larger than 1500 m3, even for the lowest critical density and longest bottleneck
duration considered (Fig. 2). This size is exceeded by 97.5% of the H. shoshone habitat
in the Sierra Nevada (Fig. 1). At the median habitat size, >99% of heterozygosity is
retained for all combinations of critical density and bottleneck duration. Not surprisingly,
similar retention percentages for allelic richness require much larger habitat size (Fig. 3).

What is important to note is that even our most extreme scenario, consisting of a ten

45

Minimum effective population

 

2.4size 5.0x102.0x102103 104 5.0x104
100% - - . r . . .
O
O
50%. .1 .. . , . . . ,. ..

Quantiles of lake volume
8
.,\°
C

 

 

0% o. T. . . .
10' 102 103 104 105 106 107

Lake Volume (m3)

 

Figure l. The quantile distribution of lake volume for a representative sample of H.
shoshone populations in the Sierra Nevada (n=169), and the corresponding minimum

effective population size for a critical density of 0.5/m3.

46

>v Z2385: 3:me H 9&3:

 

 

 

 

 

 

 

 

 

...............................................

 

E 23343:: 3:me H $5:

 

 

_ mega—no:
I I I I I I m mgoammoa
l..l..|.. _o mgonmaoa

 

. l

 

 

 

\V‘J‘H 201.31.. ‘- 1| II. .. ..... ._
o. \ x \ Q
~ .
I:~. . \ I
I o I... x. o I
m t- ..
B .
H o 1.. . o. I
,. _ mganwco:
_.. . : ;;.. IIIIII m mag—.883
o. f . .. .. _. . l..l..l.. _omo:ano:w o. I
o. . . . . . . . . o.
o mo .8 Go woo woo woo woo h5o o

T598: :8 ABJ

. ‘

3 mo

‘ I

Am oo 3

3:93: :8 ABJ

395 N. 41:0 9688: 8:0 om ooaooflomgor :flanowwmoaQ Ammo 8 SEE 4.29.3meme $5 3 goofing 3 3:93: mmNm Ea

“58:6: om coagoow men Q5 355:5: @3me om ob m:&<acm_m>:ww va BEBE: “~9me om m Bowing—{BM meagaow 9350:

mm :8 2533 ow mgmnmaoa @8588: mm m: 3355: among/8 woos—moo: :8. > 398: om <oEBm on B: mm 9m Samoa :3me m:

230: 2.4.0238 8:530: mmma ow: own—:5 8 0:0 330 go o:a wager” 2:0: 01:8— agmzw H o.m\::a.

47

>V 2:555: 0:me n o.o}:w E 35555 3:me H $50

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a
t
e
r
S :1
.nlc
k
.n|.a fi ...................................................................................................... .
f
0 oh I. .
n _ m0:0~m:o: _ @5985:
.m . IIIIII m «0:08:03 . . ._ . , . .. IIIIII m m0:08:o:m
m ON I . . . i. 1 I I I . ZomasonSODm ON I. .. I..I:I. Zommbmmmaonfi
p
O I I I. . l
r
P o.o . t . o.o . . .
8;; 0.85.: 80:00 51% 0.85.. 8 88 88 :88 ~88
35:8 050 33 $038” 050 ABJ

Emca o. 8:0 925:5: c». 2558 30:0 105000 0850a 0304 woos—moo: 530:8: 00 0,0558 3 85:0: 050 5a 330:8:
9550: m3 QC 558:5: @033 om o.o 59395;}:0“ va 555:5: 90:05 om m 5E<5cm_m\58. won—0:00: 9885: mm 50 2:304
om M0:0Bmo:m woos—moo: mm 8 555:5: 0mm0om<0 @8585: 050. 0:8 200 @8508 85m 90 85585: 80550 om 20m 0:. a.

23$. 550: 03.030 bone—mmo: 050 H a... ooa. 5:830: 50 H 3.0. 2559 om #08303 H mooo.

48

generation bottleneck at the critical density of 0.5/m3, results in over 90% retention of the
original alleles in roughly 65% of the sample of H. shoshone lakes (Fig. 3). Higher
critical densities rapidly increase retention, leading to the expectation of 90% retention
for 90% of populations when the critical density is 5/m3 and the bottleneck last for 10
generations (Figs. 1 and 3). 93 - 99 % of allelic richness is retained at the median habitat
size for the combination of parameters considered.

The pattern for genetic distance is similar to the pattern seen in heterozygosity
(Fig. 4). Nei’s D cannot be strictly interpreted as a percentage difference between two
populations, but in this case a value of 0.1 is roughly equal to 90% identity of alleles
between the two populations Wei 1987). For our most extreme scenario, habitats larger
than 4500 m3 result in D < 0.1. This includes over 90% of Sierra Nevada H. shoshone
habitat. Genetic distance of a median size habitat from its founding population is <0.0l.

The relationship between habitat size and FST shows populations in habitats >500
m3 have very low FST (Fig. 5). We include FST because of its popularity as a measure of
genetic differentiation, especially for researchers estimating migration from genetic data.
At the same time, it is less than ideal for our purposes due to its more restrictive, and less
applicable, assumptions of the relationship between populations. The FST shown (Fig. 5)
is that expected among populations when those populations are founded by the same
number of colonists from a single mixture of potential colonists. In addition these
populations grow to a large size in a single generation. The most we can safely draw
from this is that all the populations must be very small in order for very short bottlenecks
to substantially increase the amount of genetic variance distributed among populations.

F ST is further discussed below.

49

>V 735555 3:me H 095M 5 Z5555 3:me H $5:

 

 

 

 

 

 

      

 

 

 

 

 

 

 

 

 

3o _.o
_ mo:oB30: . _ @5930:
0.x IIIIII m mo:oB30:m 9w I.. .. .. . . IIIIII mmo3~m30=m
I l .l , Ho ma:anm30:m .. ., .. . l .l l 25338303
D o.o , o.o -.
'S I fit. . l
om f
N oh oh I..
.. V .
.............. I”I"l’l. (T'ltvn i .l”..l. .I... . i , . . .T Z . . , l
o.o . . llllllllll.llll.ll-.L 0.0 I VIIIIIIIULHFIIMIIIIIPIILHI I. I._
0:. ~80 goo mo: mooo Soc 3: aSo moo Soc Boo ~80
3:53: .33 A53 3:53: 33 ABWV

Emfia A. 960083 73? U mg m :0::_m30: 5533 m: 33 03383 :05530: 3:me $05 m. 3.055% 80:530: 5 3:35.55 33.
n 0.3V 8 335333 3 3:53: 33 53 3:530: 0m :3 3033303 C5 555:5 3:me H 0953 A5 55555 3:me M £50.

won—3:00.” 3:530: 3m 33 55:2 0m $535303 cog—:30: mm m: 555:5 03.333 cog—030: 33.

 

0.5

 

 

 

 

 

I '1
04 .. . i a i , t , . a, , , , ,,,,,,,
I 4
0.3 - . _ . 3
— — — Mmlmum denSIty 0.5/m
F . . . 3
ST l —— Mlnrmum den31ty S/m
0.2 T! ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
\
. L 4
\
0.1.---,\ ,,,,,,,,,,, . ,,,,,,,,,,,,,, ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, , ...... ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
\
,,,\~‘;~U, ,, ,,H,,,.UHWM
-————————-_--__---
0.0 -¥F-—-— . i
64 500 1000 1500

Habitat size

Figure 5. Expected Fsr for a group of identical populations founded at the critical density

from a single source population as determined by habitat size for minimum density =

0.5/m3 and 5/m3. See Wade and McCauley (1988).

51

DISCUSSION

Our analyses show that habitat size has a very large influence on the potential for
loss of genetic variation and increase of genetic divergence due to population bottlenecks.
A likely conservative scenario is to assume actual minimum densities are closer to 0.5/m3
and that bottleneck duration is close to the 3-4 generations suggested by past observation
(Boileau et al. 1992, McNaught et al. 1999). Under this scenario colonization/recovery
events in 70-75% of H. shoshone habitat result in <10% loss of allelic richness (Fig. 3).
Futhermore, over 90% of H. shoshone populations are resistant to significant changes in
heterozygosity or genetic distance (Fig. 2, 4). This is a hugely different outcome than
expected under the common assumption that there is no lower limit to population size.

If H. shoshone populations can often recover from lower densities, the genetic
impact of population bottlenecks will obviously be increased. Therefore, it is important
to examine the potential for bias due to our assumptions about copepod population
dynamics and genetic structure. Our estimate of critical population density is
intentionally broad, but the actual critical density could be less than 0.5/m3, because of
the potential for chemical detection in diaptomid copepods (Katona 1973, Watras 1983,
Doall et al. 1998, Sehn et al. in prep, Chapter 3) or behaviors that increase local density
of copepods (Buskey et al. 1996, Tsuda and Miller 1998). However, multiple years of
data on recovery of stocked populations in natural habitats are consistent with a critical
density within the range O.5--5/m3 (Chapter 1). At the same time, several of our
conservative assumptions about reproduction likely offset possible over-estimation of

critical density and have the potential to make populations resistant to the genetic effects

of bottlenecks in all but the smallest habitats. Namely we have assumed all eggs

52

produced are viable, hatch the next year and suffer zero mortality before reproduction.
The assumption of 100% hatching success is the most unrealistic assumption since large
numbers of eggs are known to remain dormant in the sediments. We were also extremely
conservative in presuming our reproductive males and females are able to reproduce
twice in a season. Multiple clutches are definitely common in good years (Appendix A),
but a single pair producing two clutches in a low density population should be rare
because H. shoshone must mate before each viable clutch (Watras and Haney 1980) and
the probability of mating at all is very low.

Finally, our estimates of D and loss of allelic richness are dependent on the initial
levels of variability present in the pre-bottleneck population. Increased variability results
in larger D and increased loss of allelic richness when other parameters are held constant.
For our analysis we used the same initial parameters as Nei et al. (1975) and Chakraborty
and Nei (1977) for their hypothetical fruit fly population. We thought this suitable since
we are also considering invertebrate populations with high carrying capacities and rapid
population growth. This similarity makes our results directly comparable with analyses
from Nei, which were the first to quantify the major impact that severe population
bottlenecks have on the genetics of a population. Additionally, we found no reported
values for average effective population size or mutation rate for nuclear genes in
freshwater copepods which would be more directly applicable to our calculations.

Given the limits the Allee effect should impose on the genetic outcome of
population bottlenecks and colonization events, it seems crucial to consider habitat size in
making inferences from genetic data. For example, genetic information has been used to

infer population structure and phylogeography in arctic copepods (Boileau and Hebert

53

1988, Boileau et al. 1992). These studies found significant genetic differentiation among
nearby ponds, suggesting very low migration, which seemed inconsistent with the ponds
proximity (Boileau and Hebert 1988). This paper and subsequent analyses (Boileau and
Hebert 1991 , Boileau et al. 1992) led the authors to suggest a model in which very rapid
initial growth rates--carrying capacity is reached in 5 generations or less--retain genetic
differences resulting from genetic drift due to founder events, despite migration between
populations.

Specifically, measured FST values suggested 5 or fewer individuals founded most
of the populations (Boileau and Hebert 1992). The habitats surveyed ranged from large
tundra ponds (volume = 20166 i 9882 m3) to small rock pools (volume < 50 m3)
(Boileau and Hebert 1988). The Allee effect seems to make it highly improbable that 5
individuals could successfully colonize the large ponds. In fact, measured gene frequency
divergence for the large tundra ponds were consistently lower than for populations in
much smaller rock pools in the same species (Boileau and Hebert 1988), consistent with
less severe bottlenecks in the larger habitats. Our results highlight the fact that inclusion
of very small habitats prone to severe bottlenecks could bias estimates of metapopulation
wide estimates such as FST. We suggest that the use of genetic statistics, and FST in
particular, to estimate population size during colonization events in species subject to the
Allee effect should acknowledge the ecological limitations on the populations.

Boileau and Hebert (1991) also bring up an important additional force,
colonization pattern, which interacts with the number of colonists to create the observed
population genetics. They propose that Arctic habitats were serially colonized and genetic

distance built up as a small number of colonists from a previously colonized habitat

54

successively filled each empty habitat (Boileau and Hebert 1991). The successive
colonization events proposed by the authors would act very strongly to increase
differentiation (Wade and McCauley 1988), especially in areas dominated by small pools.
In the terms of our analysis, bottleneck duration is effectively extended by each
subsequent colonization event, which could significantly increase genetic differentiation
even when habitat is dominated by large lakes.

The potential of the Allee effect to maintain genetic variation and minimize
differentiation has implications for both ecology and evolutionary biology. As in the
arctic copepod example above, population genetics are being increasingly used to try to
infer ecological processes such as migration and historical population size. This example
underscores the need for studies to directly consider minimum population size when
analyzing genetic variation in species, such as copepods, thought to experience the Allee
effect. Because the minimum effective population size can be dependent to a great
degree on habitat size, bias will result if the distribution of habitat sizes sampled does not
match the overall distribution of habitat size. The Allee effect creates differences in
population genetic expectations that are dependent on ecology, and recognizing the
potential impact of the Allee effect should allow ecologists to better understand their
results and the ways in which model assumptions are violated in their analyses.
Hopefully this will improve the accuracy of the inferences made using population
genetics.

For evolutionary biologists, a relatively large lower constraint on population size
in a species could reduce the likelihood of speciation according to some of the models

that depend on genetic reorganization due to small population size and genetic drift

55

(Wright 1931, Mayr 1963, Templeton 1980). A higher mean population size will
decrease the influence of drift on the population’s genetic makeup, presumably increasing
the influence of natural selection. Fewer bottlenecks will also preserve more genetic
variability on which natural selection can act. Additionally, in several cases very low
heterozygosity in wild populations is attributed to severe historic bottlenecks (Nei et al.
1975, O’Brien 1994). Evidence for the Allee effect, and therefore larger critical
population size, in such species with very low genetic variability suggests other causes
may need to be considered. However, it is also possible that the observed populations of
species subject to the Allee effect, such as the arctic copepods discussed above, could be
the result of successful colonization by an unexpectedly small population, following the
stochastic success of one of many repeated dispersal events with very low individual
probabilities of success.

Our analysis suggests habitat size places a crucial role in population genetics of
H. shoshone due to the Allee effect. Similar effects are possible in other species with
minimum population size determined by the Allee effect, such as conch (Stoner and Ray—
Culp 2000), and several insects (Hopper and Roush 1993, Berggren 2001, Liebhold and
Bascompte 2003). We are not suggesting a species or population becomes immune to
genetic drift or loss of genetic variability. Species with very large population size can
harbor enormous genetic variability, and a reduction of population size, even if to a large
absolute size, will cause some of that variability to be lost and increase the influence of
genetic drift (Bucklin and Wiebe 1998). Rather, this paper proposes that Allee effects,
caused by mechanisms such as mate limitation, which impose limits on minimum

population size, will have measurable and important results for the genetic dynamics of

56

the species subject to them and that this possibility, to our knowledge, is effectively
ignored in the current literature. Our results reinforce the fact the even moderate limits
on minimum population size lead to significantly different genetic outcomes than those
seen when limits on bottleneck size are ignored.

ACKNOWLEDGMENTS
This work was supported by a National Science Foundation Graduate Research
Fellowship and National Science Foundation grants DEB-9629473 and DEB-
0075509.We thank R. Knapp for the data on H. shoshone habitat sizes in the Sierra
Nevada and assistance with fieldwork, P. Soranno for the data on lake size in a large
number of North American lakes, S. Peacor for his advice on the allelic richness
simulation model, and R. Knapp, K. Scribner and G. Mittelbach for valuable suggestions

on earlier versions of the manuscript.

57

LITERATURE CITED

Allee, W. C., A. Emerson, et al. 1949. Principles of animal ecology. Philadelphia,
Saunders.

Berggren, A. 2001. Colonization success in Roesel’s bush—cricket Metrioptera roeseli:
the effects of propagule size. Ecology 82:274-280.

Boileau, M. G., and P. D. N. Hebert. 1988. Genetic differentiation of fresh-water pond
copepods at Arctic sites. Hydrobiologia 1672393-400.

Boileau, M. G. and P. D. N. Hebert. 1991. Genetic consequences of passive dispersal in
pond-dwelling copepods. Evolution 45: 721-733.

Boileau, M. G., P. D. N. Hebert, and S. S. Schwartz. 1992. Non-equilibrium gene
frequency divergence: persistent founder effects in natural populations. Journal of
Evolutionary Biology 5:25-39.

Buskey, E. J ., J. 0. Peterson, and J. W. Ambler. 1996. The swarming behavior of the
copepod Dioithona oculata: In situ and laboratory studies. Limnology and
Oceanography. 41 : 513—521 .

Chakrabory, R. and M. Nei. 1977. Bottleneck effects on average heterozygosity and
genetic distance with stepwise mutation model. Evolution 31 :347-356.

Courchamp, F., T. Clutton-Brock, and B. Grenfell. 1999. Inverse density dependence and
the Allee effect. Trends in Ecology and Evolution 14:405-410.

Doall, M. H., S. P. Colin SP, J. R. Strickler, and J. Yen. 1998. Locating a mate in 3D:
the case of T emora longicornis. Philosophical Transactions of the Royal Society B:

Biological Sciences 353:681-689.

Forsyth, S. A. 2003. Density-dependent seed set in the Haleakala silversword: evidence
for an Allee effect. 0ecologia 136:551-557.

Gascoigne, J. A., and R. N. Lipcius. 2004. Allee effects in marine systems. Marine
Ecology Progress Series 269:49-59.

Gerritsen, J. 1980. Sex and parthenogenesis in sparse populations. American Naturalist
115:718—742.

Glenn, T. C., W. Stephan, and M. J. Braun. 1999. Effects of a population bottleneck on
Whooping Crane mitochondrial DNA variation. Conservation Biology 13:1097—1107.

Katona, S. K. 1973. Evidence for sex pheromones in planktonic copepods. Limnology
and Oceanography 18:574-583.

58

Kuussaari, M., I. Saccheri, M. Camara, and I. Hanski. 1998. Allee effect and population
dynamics in the Glanville fiitillary butterfly. Oikos 82:384-392.

Lamont, B. B., P. G. L. Klinkhamer, and E. T. F. Witkowski. 1993. Population
fragmentation may reduce fertility to zero in Banksia goodii: a demonstration of the
Allee effect. 0ecologia 94:446-450.

Leberg, P. L. 2002 Estimating allelic richness: effects of sample size and bottlenecks.
Molecular Ecology 11:2445-2449.

Liebhold A. and J. Bascompte. 2003. The Allee effect, stochastic dynamics and the
eradication of alien species. Ecology Letters 6: 1 33-140.

Le Page, S. L., R. A. Livermore, D. W. Cooper, and A. C. Taylor. 2000. Genetic analysis
of a documented population bottleneck: introduced Bennett's wallabies (Macropus
rufogriseus rufogriseus) in New Zealand. Molecular Ecology 9: 753-763.

Leung, B., J. M. Drake, and D. M. Lodge. 2004. Predicting invasions: propagule pressure
and the gravity of Allee effects. Ecology 85:1651-1660.

Lewis, M. A., and P. Kareiva. 1993. Allee dynamics and the spread of invading
organisms. Theoretical Population Biology 43: 141-158.

Lierrnann, M., and R. Hilbom. 1997. Depensation in fish stocks: a hierarchic Bayesian
meta-analysis. Canadian Journal of Fisheries and Aquatic Sciences 54:1976-1984.

Mayr, E. (1963). Animal species and evolution. Cambridge, Mass, Harvard University
Press.

Maruyama, T., and P. A. F uerst. 1985. Population bottlenecks and nonequilibrium
models in population genetics. 11. Number of alleles in a small population that was
formed by a recent bottleneck. Genetics 111:675-689.

McNaught, A. S., D. W. Schindler, B. R. Parker, A. J. Paul, R. S. Anderson, D. B.
Donald, and M. Agbeti. 1999. Restoration of the food web of an alpine lake following
fish stocking. Limnology and Oceanography 44:127—136.

Myers, R. A., N. J. Barrowman, J. A. Hutchings, and A. A. Rosenberg. 1995. Population
dynamics of exploited fish stocks at low population levels. Science 269:1 106-1108.

Nei, M. 1987. Molecular evolutionary genetics. New York, Columbia University Press.

Nei, M., T. Maruyama, and R. Chakraborty. 1975. The bottleneck effect and genetic
variability in populations. Evolution 29: 1-10.

59

O'Brien, S. J. (1994). "A role for molecular genetics in biological conservation."
Proceedings of the National Academy of Sciences 91: 5748-5 755.

Odum, P. E. 1959. Fundamentals of Ecology. Philadelphia, Saunders.

Pannell, J. R. and B. Charlesworth. 1999. Neutral genetic diversity in a metapopulation
with recurrent local extinction and recolonization. Evolution 53:664-676.

Parker, B. R., F. M. Wilhelm, and D. W. Schindler. 1996. Recovery of
Hesperodiaptomus arcticus populations from diapausing eggs following elimination

by stocked salmonids. Canadian Journal of Zoology-Revue Canadienne de Zoologie
74:1292-1297.

Ramstad, K. M., C. A. Woody, G. K. Sage, and F. W. Allendorf. 2004. Founding events
influence genetic population structure of sockeye salmon (Oncorhynchus nerka) in
Lake Clark, Alaska. Molecular Ecology 13:277-290.

Robichaux, R. H., E. A. Friar, and D. W. Mount. 1997. Molecular genetic consequences
of a population bottleneck associated with reintroduction of the Mauna Kea

silversword (Argvroxiphium sandwicense ssp. sandwicense [Asteraceae]).
Conservation Biology 11:1140-1146.

Samelle, O., and R. A. Knapp. 2004. Zooplankton recovery after fish removal:
Limitations of the egg bank. Limnology and Oceanography 49: 1382-1392.

Stoner, A. W. and M. Ray-Culp. 2000. Evidence for Allee effects in an over-harvested

marine gastropod: density-dependent mating and egg production. Marine Ecology-
Progress Series 202:297-302.

Templeton, A. R. (1980). "Modes of speciation and inferences based on genetic
distances." Evolution 34: 719-729.

Tsuda, A., and C. B. Miller. 1998. Mate-finding behaviour in Calanus marshallae Frost.
Philosophical Transactions of the Royal Society B: Biological Sciences 353:713—720.

Veit, R. R., and M. A. Lewis. 1996. Dispersal, population growth, and the Allee effect:

dynamics of the house finch invasion of eastern North America. American Naturalist
1482255-274.

Wade, M. J ., and D. E. McCauley. 1988. Extinction and recolonization: Their effects on
the genetic differentiation of local populations. Evolution 42:995-1005.

Watras, C. J. and J. F. Haney. 1980. Oscillations in the reproductive condition of

Diaptomus leptopus (Copepoda: Calanoida) and their relation to rates of egg-clutch
production. 0ecologia 45:94-1 03.

60

Watras, C. J. 1983. Mate location by diaptomid copepods. Journal of Plankton Research
5:417-423.

Wright, S. 1931. Evolution in mendelian populations. Genetics 16:97-159.

Wright, S. 1938. Size of population and breeding structure in relation to evolution.
Science 87:430-431.

Wright, S. 1943. Isolation by distance. Genetics 28:114-138.

61

CHAPTER 3

THE EFFECT OF MATING BEHAVIOR AND TEMPERATURE VARIATION

ON THE CRITICAL POPULATION DENSITY OF A FRESHWATER COPEPOD
ABSTRACT

Population growth rates of dieocious zooplankton depend on the encounter rate of
potential mates, resulting in a predicted critical density for population establishment and
persistence. Existing evidence confirms a critical density for the calanoid copepod,
Hesperodiatomus shoshone, but estimates of the critical density span an order of
magnitude. We improve the precision and accuracy of previous estimates of H. shoshone
critical density by combining mating behavior from 3-dimensional video analysis with
life history data from natural populations. We also investigate the impact of temperature
variation on mating behavior and sensitivity of critical density to changes in mating
behavior vs. changes in fecundity and mortality. We estimate H. shoshone critical
density to be 0.56 — 1.3 m'3 and to be highly dependent on body size, primarily due to
changes in swimming speed. H. shoshone swimming speed increased >25% as
temperature increased from 5°C to 16°C, a temperature range relevant to their alpine lake
habitat. The corresponding decrease in critical density was roughly equivalent to the
decrease resulting from the six-fold variation in net reproductive rate observed for large-
bodied alpine copepod species. The average swimming speeds measured for H. shoshone
(1.7 - 2.4 cm/s) are dramatically faster than those previously reported for similar sized
calanoid copepods. Rapid swimming and the ability to follow pheromone trails greatly
improve the ability of H. shoshone to find mates, and the relationship between

temperature and critical density suggests that recovery or colonization events may be

more likely to succeed in warmer lakes and/or warmer years.

62

INTRODUCTION

The stocking of non-native trout species in the alpine lakes of North America has
had dramatic effects on the biotic community of these previously fishless ecosystems
(Bradford 1998, Parker 2001, Knapp et al. 2001). In the Sierra Nevada, for example,
trout stocking typically eliminates several benthic invertebrates, and the two dominant
zooplankton species, the cladoceran Daphnia melanica and the copepod
Hesperodiaptomus shoshone (Knapp et a1. 2001). If stocking is halted and the fish die-
out or are removed, H. shoshone often fails to recover despite the recovery of benthic
invertebrates and of D. melanica (Samelle and Knapp 2004). Inability to recover was also
observed with a related species under a similar scenario in the Rocky Mountains (Parker
et al. 1996). Lack of recovery is surprising given the widespread distribution of these
species (Knapp et al. 2001) and the persistence of long-lived diapausing eggs in lake
sediments (Hairston 1996, Parker 1996).

One possible explanation for the reduced resiliency of large copepods relative to
other alpine lake taxa is the difficulty of finding a mate. In a 3-dimensional habitat that is
large relative to copepod body size, the rate of male-female encounter can have an
important impact on population growth and persistence (Gerritsen 1980, Buskey 1998,
Strickler 1998). Mechanistic models based on population growth rate and encounter rate
predict a critical density for copepod population viability (Gerritsen 1980, Kiorboe 2006).
Below this threshold, mate limitation results in a negative population grth rate. This
critical density has been hypothesized to explain observed population densities (Kiorboe
2006) and the failure of H. shoshone populations to recover following extirpation by

non-native trout (Samelle and Knapp 2004).

63

Using Gerritsen’s (1980) diffusion model of encounter rate and very limited
information about swimming speed, detection distance, and life history parameters, the
critical density for H. shoshone recovery was estimated to be 0.5—5m”3 (Samelle and
Knapp 2004). Results of a recent multi-lake reintroduction experiment suggested that the
critical density was somewhat lower, between 0.2 and 3m'3 (Chapter 1). Here our aim is
two-fold, to use data on H. shoshone mating behavior to obtain a more precise and
accurate estimate of critical density, and to assess how environmental variation affects
the critical density of H. shoshone.

Improving our estimate of critical density depends on obtaining a quantitative
understanding of H. shoshone mating behavior. Various adaptations to increase
encounter rate have been identified in copepods, including swarming (Ambler et al.
1996), pheromone clouds (Nihongi et al. 2004, Kiorboe et al. 2005) and pheromone trails
(Doall et al. 1998, Tsuda and Miller 1998, Yen et al. 1998, Bagoien and Kiorboe 2005).
Earlier studies on freshwater copepods have not found evidence of pheromone trails (van
Leeuwen and Maly 1991, Nihongi et al. 2004), but recent observations by our group
indicate that H. shoshone males use pheromone trails to locate females (Sehn et al. in
prep). Theory suggests that trail-following behavior should be more common in larger
species, such as H. shoshone, and results in higher encounter rates than hydromechanical
or pheromone cloud methods of detection (Kiorboe and Bagoien 2005). Here we analyze
a large number of H. shoshone mating interactions and use the results to estimate critical
density with a model incorporating pheromone trails (Kiorboe and Bagoien 2005).

A second objective was to examine whether environmental variation affects

encounter rate and critical density of H. shoshone. We focus specifically on the role of

64

temperature in this paper. Temperature has been shown to affect many aspects of
copepod biology, including reproductive rates (Watras 1983, Williamson and Butler
1987, Chow-Fraser and Maly 1991), developmental rates (Roberston et al. 1974, Hart and
McLaren 1978), and life span (J ersabek and Schabetsberger 1995, Hirst and Kiorboe
2002). Temperature may affect encounter rate both via physiological mechanisms, such
as changes in activity or pheromone production, and physical mechanisms, including
changes in viscosity and diffusion rates. For example, Podolsky and Emlet (1993) found
that 40% of the decrease in swimming speed of sand dollar larvae (Dendraster
excentricus) over a declining temperature gradient was due to increases in viscosity.
Therefore, it is important to consider how temperature variation is likely to alter the
critical density of copepod populations, and whether significant differences can arise
solely from changes in encounter rate.
METHODS

Animal collection and maintenance

Adult H. shoshone were collected from an alpine lake (Name Lake) in the
Sawtooth range, Utah, USA on August 25, 2006. Animals were shipped overnight in
bottles on ice, to the Georgia Institute of Technology in Atlanta, GA. Animals were
maintained as a mixture of males and females in bottled spring water at 12°C and fed
rotifers. Mortality was <5% d'l and surviving animals remained healthy and active for >3
weeks. Mating activity was observed throughout this time period. Two days prior to
experiments, randomly-selected males and gravid females were separated and acclimated
to experimental temperatures.

Temperature experiments

65

Observations of mating behavior were conducted at 5, 12, and 16°C. Experiments
took place in a 3.0 L vessel containing spring water, maintained at the treatment
temperature by circulating distilled water in a large water jacket through a refrigeration
unit. Mating behavior was recorded using three—dimensional Schlieren laser videography
as developed by Strickler (Strickler & Hwang 1998) and further described by Doall
(Doall et al. 1998).

Experiments were initiated by adding males and females to the vessel at a nearly
1:1 ratio (Table 1). We observed mating in mixed groups of individuals in a relatively
large vessel because this method better mimics nature compared to interactions between a
single male and female in a more restricted container. Mating interactions at each
temperature were recorded for 4 hour periods on two consecutive days, August 5th and
6th (8 hours total). Males and females were separated between trials, maintained at their
acclimation temperature and fed rotifers. To keep sex ratio and density as constant as
possible, dead animals were replaced with randomly selected replacement individuals of
the appropriate sex that had been acclimated to the treatment temperature for > 24 hours
(Table 1).

Video recordings were processed to identify H. shoshone mating interactions. We
defined mating events as behavior culminating in one individual being clasped by a
pursuing individual, followed by the pair entering a characteristic tumbling behavior.

The sex of animals was not obvious from the recording, so the pursuer was assumed to be
a male and the pursued a female. Previous observations indicated that males
preferentially pursue females (Yen, unpublished data). We digitized several seconds of

swimming at 1/30 sec intervals prior to each attempted mating for which both animals

66

Table 1: The conditions and results of mating trials. Density and sex ratio varied based
upon mortality. The same individuals were used in trial 1 and 2, except for replacements
due to mortality between trials. Observed mating attempts are the total seen in the
recording of the 4 hour trial and analyzable encounters are the subset from which .

complete 3D positions were obtainable.

 

 

Temperature Density (# L'I) Males : Observed Analyzable
Females mating attempts encounters
5 Trial 1 8 11 :12 44 14
Trial 2 7 10 : 11 21 11
12 Trial 1 7 10:11 27 10
Trial 2 7 10: ll 31 13
16 Trial 1 7 9 : 8 77 22
Trial 2 7 10: 10 106 35

 

67

were visible in both the x-z and y-z views. This resulted in 105 analyzable events (Table
1).

We quantified the swimming speeds of the animals prior to detection and the
length of the trail followed by the male (see Doall et al. 1998 and Sehn et al. in prep for
calculations) from the digitized video. Initiation of trail following was recognized by an
abrupt change in direction and noticeable acceleration, and confirmed by visualization of
the digitized male and female paths. Trail length was estimated as the length of the
female's swimming path between her and the male at the point pursuit began. This was
defined as the point along the female path closest to the male when he reacted to the
presence of the trail. Maximum trail length is the most relevant measure for an encounter
model, because we are interested in how far away males are able to detect females,
whereas average trail length is determined by where the males encounter the trail, even if
they would have been able to successfully follow a longer trail. We wanted a quantitative
estimate of maximum trail length rather than the absolute limit, which may only occur
under ideal conditions or for extraordinary individuals, so we arbitrarily took the average
of the 90th percentile across all treatments and used this as our estimate of L in the
encounter model (Equation 1 below).

ANOVA was used to evaluate the effect of temperature on swimming speed,
average trail length and average trail age. There were no significant differences in these
parameters between experiments performed on different days (P > 0.3), so we pooled
observations across days. Post—hoc pairwise comparisons were done using Tukey-
Kramer HSD (SYSTAT version 9).

Encounter model and estimate of critical density

68

We estimated the search volume rate (B, m3 day") for H. shoshone using an

equation for a cruising male and a female pheromone trail (Bagoien and Kiorboe 2005):

fl:2L113/)["‘D_:'L_+S] (1)

where S is the sensory reach of the copepod (the length of one antennule, here roughly
equivalent to the body length), L is the length of the pheromone trail, DP is the diffusion
coefficient of the pheromone, v is the swimming speed of the female prior to detection,
and u 20 is the two-dimensional component of male swimming speed prior to pursuit. DP
was assumed to be 105, typical of small biological molecules (Yen et al. 1998), and two-
dimensional component of velocity was estimated based on random 3-dimensional
swimming (Kiorboe and Bagoien 2005).

The search volume rate was then used to estimate critical density (NC) following
Genisten (1980). We chose this formulation because we were able to estimate R0 with
existing data, and could not accurately estimate the stage-specific mortalities required by
the alternate formulation (Kiorboe 2006). The critical sexual encounter rate (Z)

necessary to maintain a population without growth is (Gerritsen 1980):

Z = —_1 In [M], (2)
t R0

where t is the is the length of the breeding season and R0 is the net reproductive rate.
We replaced the spherical, diffusion model of encounter (Gerritsen 1980) with the

estimate based on the detection of pheromone trails. Assuming a 1:1 sex ratio, we obtain:

_21n[£3;1]
N. 0 . (3)

 

69

Values for annual rate of increase from small population size were estimated from
the first 4 years of successful reintroductions of H. shoshone (r = 0.236, from Chapter 1)
and Hesperodiaptomus arcticus (r = 1.57, McNaught et al. 1999). These reintroductions
offer a unique opportunity to directly assess net population growth rate since abundances
were below equilibrium. H. shoshone are univoltine and encounter rate is per day, so R0
= er/m. Maximum length of breeding season is assumed to be 60 days based on
extensive sampling of Sierra Nevada H. shoshone populations from 2002-2006 (Kramer
unpublished data) and estimated adult mortality rates (Chapter 1).

RESULTS

Temperature had a significant effect on the swimming speed prior to pursuit of
both males and females (ANOVA, p < 0.001, Figure 1). Male swimming speeds were
higher at 16°C (p<0.001), and not different between 5°C and 12°C. (Figure 1). Female
swimming speeds were also indistinguishable between 5°C and 12°C, and at 16°C were
significantly higher than 12°C (p<0.001) and nearly significantly different than 5°C
(p=0.06). Average trail lengths varied from 1.5 cm at 5 to 1.2 cm at 16°C, but there was
no significant effect of temperature (P > 0.2 ). We estimated maximum trail length to be
L = 3.14 cm.

Given our estimates of trail length and swimming speed (Fig. 1) and our range of
estimates of growth rate (see Methods), we estimated a range of critical densities, 0.56 —
0.98 m-3, for a population of 3mm long H. shoshone (Fig. 2). Average body length in
many Sierra Nevada populations is 2.5mm (Appendix A), causing decreases in

swimming speed and anntenule length; using swimming speeds at 12°C from a small

70

 

3.0 f I I.

a female

25_ I male _

 

 

Velocity (cm/s)
N
‘?

 

 

 

 

 

 

 

 

1.0
5°C 12°C 16°C

Treatment

Figure 1. Average pre-pursuit swimming velocity (cm/s) for female and male H.
shoshone. Sample size is indicated inside each bar and error bars represent the standard
deviation. According to ANOVA results, treatment temperature had a significant effect

on both female (p < 0.001) and male (p<0.001) velocity.

71

 

.0 .O .O 7"
N 00 (D O
I I I

Critical Density (# m’3)

.0
C)
I

 

 

 

 

 

 

0.5
5 12 16

Temperature (C)

Figure 2. Critical density calculated using male and female velocity from each
temperature treatment and low (1 .00065 day") or high (1.0043 day") net reproductive

rate. Breeding season length is 60 d.

72

sample of Sierra Nevada animals (Sehn et al. in prep) the resulting estimate of critical
density is 0.86 — 1.30 m.

An increase in swimming speed associated with an increase in temperature from 5
to 16°C decreases critical density nearly as much as the difference across the range of
observed annual rates of increase (Fig. 2). In order to further assess the relative
importance of variation in parameters influencing encounter rate vs. variation in
reproductive rate we looked at the sensitivity of critical density to swimming speed, trail
length and population growth rate. Proportionally equivalent changes in swimming speed
and trail length had very similar impacts on critical density (Fig. 3b, 0). Conversely,
changes in population growth rate had little influence on critical density (Fig. 3a). It is
also apparent that proportional decreases have a larger absolute effect than increasing the
parameters by the same proportion (Fig. 3).

DISCUSSION

Our behavioral data enabled estimation of a range of critical density for H.
shoshone in the Sierra Nevada (0.86 - 1.3 m3) that was more precise than in a previous
paper (0.5 — 5 m3, Samelle and Knapp 2004), and congruent with the results of a multi-
lake reintroduction experiment (0.2 — 3 m3, Chapter 1). We also found that increased
swimming speeds at higher temperatures (Fig. 1) have a relatively large effect on
estimates of critical density (Fig 2). Model calculations suggest that critical density is
more sensitive to temperature-induced changes in encounter rate than to changes in
reproductive rate (Fig. 3) over the range of temperatures found in H. shoshone habitat.

Because both encounter rate and life history parameters are expected to vary with

73

 

4 ' a) Annual population growth rate

Critical Density (3 m'3)
N

 

0
4 ' b) Swimming speed

Critical Density (3 m3)
N

 

4 ' c)Trail length

Critical Density (3 ma)
N

 

 

 

- O 50 1 0

Proportional change
Figure 3. Sensitivity of estimated critical density to proportional changes in: a) annual
population growth rate, b) swimming velocity, and c) trail length. Default value is

velocity and trail length for males and females at 12°C (see text), 60 day breeding season,

and net reproductive rate = 1.00065 d".

74

temperature, critical density in this species likely varies substantially among lakes and
between years.

Much of the increase in the precision of the critical density estimate is due to our
use of field-measured, H. shoshone-specific life history parameters, however we
introduce variation due to swimming speed that has not been considered in previous
estimates of critical density. We have also increased the accuracy of the critical density
estimate by using a realistic model of mating behavior in combination with
experimentally measured parameters.

In addition to expected improvements in our estimates of critical density, this
experiment provided evidence that temperature strongly influences the critical density of
copepod populations by affecting swimming speed (Fig 2). Notably, male swimming
speed tended to increase more than female speed across our temperature gradient (Fig. 1),
and increases in male speed have a much larger effect on critical density than equivalent
increases in female speed (Eq. 1). Whether temperature affects pheromone trails is
unclear from these results. To accurately test L, the trail length relevant to encounter rate,
a sample size large enough to ensure sufficient sampling of the longest trails is necessary,
something we didn’t achieve at 5 or 12°C. It is interesting to note that we observed a
significantly higher number of mating events (ANOVA, p=0.043) in trials at 16°C (Table
1) and this increase exceeds that predicted by differences in swimming speed alone,
implying that perhaps some other factors related to temperature were affecting either
encounter rate or another aspect related to the initiation of mating behavior.
Alternatively, animals may have been distributed closer to the bottom or sides of the

chamber at lower temperature, reducing our ability to observe their activity.

75

Increased water temperature reduces critical density by increasing encounter rates,
but can also affect at least one other component of critical density, namely reproductive
rate. In general, warmer water results in increased egg production rate by copepods
(Hirst and Kiorboe 2002). In one alpine copepod species, egg production rate was
maximized at 20°C, while another species experienced highest fecundity at an
intermediate (10°C) temperature (J ersabek and Schabetsberger 1995). However,
reproductive rates appear to have little impact on critical density relative to other factors.

Temperature could also have an effect on breeding season length and proportional
increases in breeding season length have effects similar to increases in swimming speed
and trail length for the relevant range of parameter values (Equation 3).! In warm years
alpine lakes will experience longer ice-free periods, presumably allowing adults to appear
earlier in the year and likely improving food quality and/or quantity, but increased
temperature also results in increased mortality rates (Hirst and Kiorboe 2002). Because
of these opposing influences, we used a single estimate of breeding adult presence in the
lake, 60 days, which is equal to or greater than the maximum observed in 5 years of
sampling Sierra Nevada H. shoshone populations, and is conservative with respect to the
lower bound of the critical density.

Combining the effect of temperature on mating behavior and other components of
critical density in H. shoshone, it seems likely that warmer temperatures result in lower
critical density for these alpine copepods. Mean temperatures for the ice-free period of
seven Sierra Nevada lakes (3358 — 3583 m elevation) measured from 1996-2002 ranged
from 6.4 — 16.5°C (Knapp and Samelle unpublished data). This range is nearly identical

to that examined in this experiment, and our results suggest ~25% reduction in critical

76

density at higher temperatures, due to increased swimming speed. A 25% decrease in
critical density is a large change when the absolute number of individuals involved is
considered. The median volume of lakes with established populations of H. shoshone in
the Sierra Nevada is 1.3 * 105 m3 (Chapter 2). A decrease in critical density from 1 m"3
to 0.75 m'3 is a decrease of 33,000 animals that must initially colonize in order forthe
population to become established. It follows that recovery or colonization events taking
place in warmer lakes or in warmer years may be more likely to succeed in establishing a
population. The results also highlight the importance of body size. Populations of H.
shoshone with body size closer to 2.5 mm than to the 3 mm animals used here are likely
to have significantly lower swimming speeds (see Sehn et al. in prep) and, therefore, a
higher critical density.

More generally, these results suggest H. shoshone deals with the potential challenges of
finding mates through two important adaptations, the ability to produce and follow
pheromone trails (Sehn et al in prep) and rapid swimming. The swimming velocity of H.
shoshone is much higher than published swimming speeds for other calanoid copepods
(Fig. 4), and several times higher than the similarly sized marine copepod Calanus
marshallae (Tsuda and Miller 1998). A potential explanation for this difference is that H.
shoshone evolved in the absence of fish, meaning adults have no pelagic predators, and
many of the lakes also lack potential predators of juveniles, such as cyclopoid copepods
and Chaoborus (Samelle and Knapp 2004). Increases in swimming speed increase the
encounter rate with predators in much the same way as they increase the encounter rate
with mates (Gerritsen 1980) and the lack of predators may have removed selection

against rapid swimming in these large animals, or at least reduced the

77

 

3 l l l l l
o H. shoshone
A o Other calanoids A
(I)
E
3 2 — —
.0 D
o
o
O.
U)
U)
.E 0
3
a) o
. C
O
O
0 I l I I I

 

 

 

0.5 1.0 1.5 2.0 2.5 3.0 3.5
Body size (mm)

Figure 4. Swimming speed as a function of body size for H shoshone (open shapes) and
six other calanoid copepod species (filled circles). Measured swimming speeds are
shown for Sierra Nevada H. shoshone at 12°C (O), and Utah H. shoshone at 12 (D) and
16°C (A). Data from Tsuda and Miller (2004), Doall et al. (1998), Nihongi et al. (2004),

Kiorboe and Bagoien (2005), and Sehn et al. (in prep).

78

cost of large size, allowing rapid swimming. Large body size is expected to favor
pheromone production, due to metabolic constraints, and existing data on copepod mate
search capability suggest larger species are more likely to produce pheromones (Kiorboe
and Bageien 2005). Adaptations favoring low critical density are important to an animal
that must colonize and persist in isolated habitats, especially because calaniod copepods
have limited dispersal capabilities (Jenkins and Buikema 1998, Caceres and Soluk 2002).

This study suggests that when densities are low, behavioral components of mating
are equal to or more important in H. shoshone population grth than factors influencing
egg production and mortality. It seems likely that other large freshwater copepods, such
as H. arcticus, will have comparable mate search capabilities due to similarities in
habitats and body size. Additionally, the relationship between temperature, swimming
speed and encounter rate suggests a pathway through which temperature changes could
affect reproductive rate in copepods at densities near the critical density.

ACKNOWLDEGMENTS

We thank Ryan Lockwood for providing live H. shoshone, Rachel Lasley for help
with mating experiments and Megan Heaphy and Jennifer Sehn for help with image and
data analysis. Kevin Pangle provided comments on an early draft of this manuscript.
The work in the Sierra Nevada would not have been possible without the help of Roland
Knapp, and was financially supported by a National Science Foundation Graduate
Research Fellowship, National Science Foundation grants DEB-9629473, DEB-0075509,
an REU supplement, VESR Graduate Student grant, and a Sigma-Xi Grant-in-aid of
Research. Research and collecting permits were provided by the Inyo and Sierra

National Forests, and the California Department of Fish and Game.

79

LITERATURE CITED

Ambler, J. W., S. A. Broadwater, E. J. Buskey, and J. 0. Peterson. 1996. Mating behavior
in swarms of Dioithona oculata, p. 287-299. In P. H. Lenz, D. K. Hartline, J. E.
Purcell, and D. L. MacMillan [eds], Zooplankton: Sensory ecology and physiology.
Gordon and Breach.

Bagoien, E., and T. Kiorboe. 2005. Blind dating - mate finding in planktonic copepods. 1.
Tracking the pheromone trail of Centropages typicus. Marine Ecology-Progress Series
300: 105-115.

Bradford, D. F ., S. D Cooper, T. M. Jenkins, K. Kratz, O. Samelle, A. D. Brown. 1998.
Influences of natural acidity and introduced fish on faunal assemblages in California
alpine lakes. Canadian Journal of Fisheries and Aquatic Sciences 55: 2478-2491

Buskey, E. J. Components of mating behavior in planktonic copepods. Journal of Marine
Systems 15: 13-21

Caceres, C. E., and D. A. Soluk. 2002. Blowing in the wind: a field test of overland
dispersal and colonization by aquatic invertebrates. 0ecologia 131:402-408.

Chow-Fraser, P., and E. J. Maly. 1991. Factors governing clutch size in 2 species of
Diaptomus (Copepoda, Calanoida) Canadian Journal of Fisheries and Aquatic
Sciences 48: 364-370.

Doall, M. H., S. P. Colin SP, J. R. Strickler, and J. Yen. 1998. Locating a mate in 3D:
the case of T emora longicornis. Philosophical Transactions of the Royal Society B:
Biological Sciences 353:681-689.

Gerritsen, J. 1980. Sex and parthenogenesis in sparse populations. American Naturalist
115:718—742.

Hairston, N. G., Jr. 1996. Zooplankton egg banks as biotic reservoirs in changing
environments. Limnology and Oceanography 41 : 1087—1092.

Hart, R. C ., and I. A. McLaren. 1978. Temperature acclimation and other influences on
embryonic duration in the copepod Pseudocalanus sp. Marine Biology 45: 23-30

Hirst, A. G., and T. Kiorboe. Mortality of marine planktonic copepods: global rates and
patterns. Marine Ecology-Progress Series 230: 195-209

Jenkins, D. G., and A. L. Buikema. 1998. Do similar communities develop in similar

sites? A test with zooplankton structure and function. Ecological Monographs 68:421-
443.

80

J ersabek, C. D., and R. Schabetsberger. 1995. Resting egg production and oviducal
cycling in two sympatric species of alpine diaptomids (Copepoda: Calanoida) in
relation to temperature and food availability. Journal of Plankton Research 17 : 2049-
2078

Katona, S. K. 1973. Evidence for sex pheromones in planktonic copepods. Limnology
and Oceanography 18:574-583.

Kiorboe, T. 2006. Sex, sex-ratios, and the dynamics of pelagic copepod populations.
0ecologia 148: 40-50

Kiorboe, T. and E. Bagoien. 2005. Motility patterns and mate encounter rates in
planktonic copepods. Limnology and Oceanography 50:1999-2007.

Kiorboe, T., E. Bagoien, and U. H. Thygesen. 2005. Blind dating—mate finding in
pelagic copepods. II. The pheromone cloud of Pseudocalanus elongates. Marine
Ecology-Progress Series 300: 117-128.

Knapp, R. A., K. R. Matthews, and O. Samelle. 2001b. Resistance and resilience of
alpine lake fauna to fish introductions. Ecological Monographs 71:401-421.

McNaught, A. S., D. W. Schindler, B. R. Parker, A. J. Paul, R. S. Anderson, D. B.
Donald, and M. Agbeti. 1999. Restoration of the food web of an alpine lake following
fish stocking. Limnology and Oceanography 44:127-136.

Nihongi, A., S. B. Lovem, and J. R. Strickler. 2004. Mate-searching behaviors in the
freshwater calanoid copepod Leptodiaptomus ashlandi. Journal of Marine Systems
49:65-74

Parker, B. R., D. W. Schindler, D. B. Donald, R. S. Anderson. 2001. The effects of
stocking and removal of a normative salmonid on the plankton of an alpine lake.
Ecosystems 42334-345

Parker, B. R., F. M. Wilhelm, and D. W. Schindler. 1996. Recovery of
Hesperodiaptomus arcticus populations from diapausing eggs following elimination
by stocked salmonids. Canadian Journal of Zoology-Revue C anadienne de Zoologie
74:1292-1297.

Podolsky, R. D. and R. B. Emlet. 1993. Separating the effects of temperature and
viscosity on swimming and water-movement by sand dollar larvae (Dendraster-
Excentricus). Journal of Experimental Biology. 176: 207-221.

Robertson, A., Gehrs, C. W., Hardin, B. D., Hunt, G. W. 1974. Culturing and ecology of
Diaptomus clavipes and Cyclops vernalis. US. Environmental Protection Agency,
Washington DC. (EPA=660/3-74-006).

81

Samelle, O., and R. A. Knapp. 2004. Zooplankton recovery after fish removal:
limitations of the egg bank. Limnology and Oceanography 49:1382-1392.

Strickler, J. R. 1998. Observing free-swimming copepods mating. Philosophical
Transactions of the Royal Society of London Series B-Biological Sciences 353: 671-
680

Strickler, J. R. and Hwang, J .-S. 1998. Matched spatial filters in long working distance
microscopy of phase objects. In P. C. Cheng, P. P. Hwang, J. L. Wu, G. Wang and H.
Kim [eds.], Focus on multidimensional microscopy. World Scientific.

SPSS Inc. 1998. SYSTAT version 9.

Tsuda, A., and C. B. Miller. 1998. Mate-finding behaviour in Calanus marshallae Frost.
Philosophical Transactions of the Royal Society B: Biological Sciences 353:713—720.

VanLeeuwen, H. C. and E. J. Maly. 1983. Changes in swimming behavior of male
Diaptomus leptopus (Copepoda, Calanioda) in response to gravid females. Limnology
and Oceanography 36: 1188-1195.

Watras, C. J. 1983. Mate location by diaptomid copepods. Journal of Plankton Research
5:417-425.

Watras, C. J. and J. F. Haney. 1980. Oscillations in the reproductive condition of
Diaptomus leptopus (Copepoda: Calanoida) and their relation to rates of egg-clutch
production. 0ecologia 45:94-103.

Williamson, C. E., and N. M. Butler. 1987. Temperature, food and mate limitation of
copepod reproductive rates: separating the effects of multiple hypotheses. Journal of
Plankton Research 9:821 -836.

Yen, J ., M. J. Weissburg, and M. H. Doall. 1998. The fluid physics of signal perception

by mate-tracking copepods. Philosophical Transactions of the Royal Society of
London Series B-Biological Sciences 353: 787-804

82

APPENDIX A

Hesperodiaptomus shoshone is a large (>2mm) deeply-pigmented copepod that
occurs in alpine lakes in the Sierra Nevada and the Rocky Mountains from Colorado to
British Columbia (Pennak 1978). As part of a multi-year project on the recovery ability
of H. shoshone populations, I collected samples from multiple populations in the Sierra
Nevada. Some populations were sampled a single time, but others were sampled in
multiple years and multiple times per year. I also conducted a month long translocation
experiment in which H. shoshone were placed in cages in 4 lakes. Some data from these
samples is summarized here in order to improve the understanding of the life-history and

population dynamics of this species.

SAMPLE AREA

Animals were collected between 2001 and 2006 from 19 persistent populations, 5
experimentally reintroduced populations and l naturally recovering population. All lakes
were above 3290 meters in elevation and located in the Sierra National Forest, Inyo
National Forest and Kings Canyon National Park (Table 1). Persistent populations can be
categorized based on trout stocking history as: fish present, never stocked, or stocked but
now fishless (Knapp et al. 2001). Populations in the final category are expected to have
been extirpated by fish predation and to have recovered from the egg bank following fish

disappearance (Knapp et al. 2001, Knapp and Samelle in prep).

83

Table 1: Lakes sampled, with lake ID (when known), name (when in quotes the name is
not found of topographical maps), morphological data, stocking history(S-F=stocked, fish
present, NS=never stocked, S-FL=stocked previously, now fishless, S-FL, Exp=Now
fishless, experimentally reintroduced), the number and years each lake was sampled,
additional samples which provided no information to this analysis are not included.

(sampling differs for experimentally populations, see Methods).

 

 

Lake Max # of Years
ID Lake Name Elevation Area Depth Status samples sampled
40222 Ramona 3290 12.5 1 1.25 S-F l 2003
41203 "Pavilion, Lower" 3323 8.2 20 NS 1 2003
50153 Goethe, Lower 3513 4.5 19.5 S-F 6 2001-2003
50154 Goethe, Upper 3514 23.1 30 S-F 4 2001-2003
50176 Lobe, Upper 3291 2.2 8.5 S-F 1 2003
50193 Petite, Lower 3504 1.0 7 S-FL 3 2001-2003
50194 Petite, Upper 3504 0.6 6.5 NS 3 2001-2003
50207 Puppet 3422 20.7 5.3 S-F l 2003
50224 Wahoo 2 3443 3.5 13 S-FL 5 2001-2003
50423 Spire 3523 6.7 30 S-F 4 2001-2003
50424 Split 3413 1.6 20.5 S-F 5 2001-2003
52103 "Frog" 3632 0.9 5 NS 2 2003-2004
52121 "Dissertation" 3602 1.1 7 NS 8 2002-2005
52127 "Freedom" 3547 1.8 5.5 NS 1 2003
10222 “Barrett, Upper” 3554 1.1 5.2 NS 1 2004
10223 “Barrett, Middle" 3495 4.0 14.8 S-FL 1 2004
10230 “Glacier 3533 14.3 10.3 NS 1 2004
10257 “Southfork Pass“ 3587 17.8 19.5 NS 1 2004

 

84

Table 16 (cont’d)

 

 

Lake Max # of Years
ID Lake Name Elevation Area Depth Status samples sampled
10263 “Palisade, Upper” 3450 4.8 13.2 S-FL 1 2004
50183 Marmot 3583 3.03 8 S-FL 2 2004-2005
50219 Square 3443 1.71 3.5 S-FL, Exp 10 2003-2005

“No Good” 3516 1.67 5 S-FL, Exp 6 2003-2005
Knob 3358 3.39 5.5 S-FL, Exp 2 2003
50133 Cony 3492 1.43 3.3 S-FL, Exp 1 2003

 

85

METHODS

Animals were sampled using 2 or more vertical tows of a 30 cm diameter
plankton net with 64pm mesh. Tows were conducted in the deepest area of the lake from
a float tube. In the case of the experimentally reintroduced populations, samples were
collected with a 1 m diameter, 350 um mesh net. This mesh is big enough to capture egg
bundles and copepodites (Chapter 1). Samples were preserved in 90% ethanol. Three or
more replicate, lml subsamples were examined at 40X with a Sedgewick—Raf’ter
chamber, and the H. shoshone were assigned to the following categories, adult male,
copepodite male, adult female, copepodite female, sex-indetenninate copepodites, and
nauplii. Adult females were additionally assigned a reproductive condition of gravid,
gravid w/eggs, non-gravid w/eggs, or non-gravid. Observed egg bundles were measured
and the number of eggs was counted. The length of the first 50 individuals of each
category was measured. Co-occuring copepods and cladocerans were also measured and

enumerated, but that data will not be presented here.

LIFE HISTORY
Hatching
The data strongly suggest that a single generation of H. shoshone hatches
synchronously from the egg bank prior to ice-out. Observing nauplii was very rare, even
when lakes were sampled while partially ice-covered. Only one sample from a persistent
population contained adults and more than two nauplii, and only six of 47 samples
containing adults also contain 2 or 1 nauplii. Seventeen of the samples from persistent

population contain only adults. Only two samples taken after July contained nauplii.

86

Nauplii were never observed in the experimentally reintroduced populations or the
Marmot Lake population. Eggs much hatch sufficiently early for substantial
development to take place before the ice melts.

Nauplii were not observed in the caged populations, even though a large number
of eggs were produced. This suggests all the eggs produced were resting eggs.
Development

Copepodites were not identified to stage, so development times for various stages
can not be estimated. We can make some rough estimates at the population level. In six
cases sampling provided a time series taking the population from >90% copepodites to
80-100% adults. These samples were 30 days apart for Spire and Split Lakes in 2001,
and Dissertation Lake in 2003, and only 10 days (Square Lake 2003), 13 days (Marmot
Lake 2005), and 16 days apart in the other 3 cases (No Good Lake 2004). While an
estimate of development time is difficult to make from these data, it is clear that time to
maturity can be prolonged and that development is usually highly synchronous.

Adults were still present in all lakes which were sampled a month or more after
adults were first observed to be predominate (n=8). In a similar time period of four
weeks, females in the caged populations produced multiple clutches, proving females can
mate multiple times after maturity and allowing for an estimate of interclutch duration in
these animals. In these four lakes (see Chapter 1) interclutch duration varied from 13 -20
days. We can assume females have enough time to produce 2 or more clutches each
summer.

Body size

87

Average body size varied among lakes and years. Average body size among all
lakes sampled in 2003 was similar to, but slightly smaller and less variable than, all
lake*year combinations for which data was available (Fig. 1). Males tended to be smaller
than females on average (Fig. 1) and in each lake*year combination a males were 83 um
(i121) smaller than non-gravid females (paired t-test, p<0.0001, n=36) and 152 um
(i156) than gravid females (p<0.0001, n=3 8). Non-gravid females were 71 um (i133)
smaller than gravid females (p<0.003, n=36). Males and females differed less in size
within a given lake*year than sexes did between lakes. There was a large difference
between the maximum average size observed, 3200 um for gravid females in No Good
Lake (2004), and the minimum average size, 1820 um for males in Spire Lake (2001).
Sex ratio

Sex ratio was varied widely among lakes, but also varied within lakes among
samples. Average sex ratio was 0.99 (21:0.61) among 40 lake*year combinations with a
max of 2.5 males:female and minimum of 0 (two cases, fewer than 5 adults sampled).
When multiple samples were taken from a single lake in a single year, all data were
pooled to calculate the sex ratio for that lake. Average and standard deviation were
similar among lakes with multiple samples vs. lakes sampled a single time. Within a lake
sex ratio varied as much a 6—fold among samples within the same year. Further analysis
should determine if sex ratio varies temporally or is it is merely highly stochastic with

regard to a given sample, perhaps suggesting spatial variation in sex ratio.

88

 

3000

I 2003

 

  
 

 

Length (pm)
a
8

 

 

 

Males Non-gravid females Gravid females

Category

Figure 1. Average and standard deviation of body length (um) for adult H. shoshone in
2003 (n=14 lakes) and all lake*year combinations (n=40).

89

Reprodutive rate

Average clutch size was 12.3 eggs (i5.1, max = 23, min = 6)) among 29
lake* year combinations. Average clutch size for the caged populations was 16 (3:41)
(Chapter 3).

We were able to make an estimate of growth rate from the experimentally re-
introduced population in Square Lake (see Chapter 3). However, this population growth
rate is likely to be dependent both upon the density of individuals and the state of the egg
bank (Chapters 1 and 3).

Mortality
Mortality was estimated for the caged populations and the two most thoroughly

sampled populations (see Chapter 1).

90

LITERATURE CITED

Knapp, R. A., K. R. Matthews, and O. Samelle. 2001. Resistance and resilience of alpine
lake fauna to fish introductions. Ecological Monographs 71:401-421.

Pennak, R.W. 1978. Freshwater invertebrates of the United States. Wiley, New York.

91

LITERATURE CITED

Allan, J. D. 1976. Life-history patterns in zooplankton. American Naturalist 1102165-
180.

Allee, W. C., A. Emerson, et al. 1949. Principles of animal ecology. Philadelphia,
Saunders.

Amarasekare, P. 1998. Allee Effects in Metapopulation Dynamics. The American
Naturalist 152:298-302.

Ambler, J. W., S. A. Broadwater, E. J. Buskey, and J. 0. Peterson. 1996. Mating behavior
in swarms of Dioithona oculata, p. 287-299. In P. H. Lenz, D. K. Hartline, J. E.
Purcell, and D. L. MacMillan [eds], Zooplankton: Sensory ecology and physiology.
Gordon and Breach.

Bagoien, E., and T. Kierboe. 2005. Blind dating - mate finding in planktonic copepods. 1.
Tracking the pheromone trail of Centropages typicus. Marine Ecology-Progress Series
300: 105-115.

Berggren, A. 2001. Colonization success in Roesel’s bush-cricket Metrioptera roeseli:
the effects of propagule size. Ecology 82:274-280.

Boileau, M. G., and P. D. N. Hebert. 1988. Genetic differentiation of fresh-water pond
copepods at Arctic sites. Hydrobiologia 167:393-400.

Boileau, M. G. and P. D. N. Hebert. 1991. Genetic consequences of passive dispersal in
pond-dwelling copepods. Evolution 45: 721-733.

Boileau, M. G., P. D. N. Hebert, and S. S. Schwartz. 1992. Non-equilibrium gene
frequency divergence: persistent founder effects in natural populations. Journal of
Evolutionary Biology 5:25-39.

Boukal, D. S. and L. Berec. 2002. Single-species models of the Allee effect: Extinction
boundaries, sex ratios and mate encounters. Journal of Theoretical Biology 2182375-
394

Bradford, D. F., S. D Cooper, T. M. Jenkins, K. Kratz, O. Samelle, A. D. Brown. 1998.
Influences of natural acidity and introduced fish on fauna] assemblages in California
alpine lakes. Canadian Journal of Fisheries and Aquatic Sciences 55: 2478-2491

Brassil, C. E. 2001. Mean time to extinction of a metapopulation with an Allee effect.
Ecological Modelling 14319-16.

Buskey, E. J. Components of mating behavior in planktonic copepods. Journal of Marine
Systems 15: 13-21.

92

Buskey, E. J ., J. 0. Peterson, and J. W. Ambler. 1996. The swarming behavior of the
copepod Dioithona oculata: In situ and laboratory studies. Limnology and
Oceanography. 41 : 513—521 .

Caceres, C. E., and D. A. Soluk. 2002. Blowing in the wind: 3 field test of overland
dispersal and colonization by aquatic invertebrates. 0ecologia 1312402-408.

Calabrese, J. M. and W. F. Fagan. 2004. Lost in time, lonely, and single: reproductive
asynchrony and the Allee effect. American Naturalist 164:25-37.

Campbell, M. M. 1976. Colonisation of Aphytis melinus DeBach (Hymenoptera,
Aphelinidae) in Aonidiella aurantii (Mask) (Hemiptera, Coccidae) on citrus in South
Australia. Bulletin of Entomological Research 65:659-668.

Cappuccino, N. 2004. Allee effect in an invasive alien plant, pale swallow-wort
Vincetoxicum rossicum (Asclepiadaceae). Oikos 106:3-8.

Chakrabory, R. and M. Nei. 1977. Bottleneck effects on average heterozygosity and
genetic distance with stepwise mutation model. Evolution 31 :347-356.

Chow-Fraser, P., and E. J. Maly. 1991. Factors governing clutch size in 2 species of
Diaptomus (Copepoda, Calanoida) Canadian Journal of Fisheries and Aquatic
Sciences 48: 364-370.

Courchamp, F., T. Clutton-Brock, and B. Grenfell. 1999. Inverse density dependence and
the Allee effect. Trends in Ecology and Evolution 14:405-410.

Cruickshank, I., W. S. C. Gurney, and A. R. Veitch. 1999. The characteristics of
epidemics and invasions with thresholds. Theoretical Population Biology 56:279-292.

Dennis, B. 2002. Allee effects in stochastic populations. Oikos 96:389-401.

Doall, M. H., S. P. Colin SP, J. R. Strickler, and J. Yen. 1998. Locating a mate in 3D:
the case of Temora longicornis. Philosophical Transactions of the Royal Society B:
Biological Sciences 353:681-689.

Drake, J. M. and D. M. Lodge. 2006. Allee effects, propagule pressure and the
probability of establishment: Risk analysis for biological invasions. Biological
Invasions 8: 365-375.

Ellner, S. P., N. G. Hairston, C.M. Keams, and D. Babai. 1999. The roles of fluctuating
selection and long-term diapause in microevolution of diapause timing in a freshwater

copepod. Evolution 53:1 11-122.

Forsyth, S. A. 2003. Density-dependent seed set in the Haleakala silversword: evidence
for an Allee effect. 0ecologia 136:551-557.

93

Gascoigne, J. A., and R. N. Lipcius. 2004. Allee effects in marine systems. Marine
Ecology Progress Series 269249-59.

Geiling, W. T. & R. S. Campbell, 1972. The effects of temperature on the development

rate of the major life stages of Diaptomus pallidus Herrick. Limnology and
Oceanography 17: 304—307.

Gerritsen, J. 1980. Sex and parthenogenesis in sparse populations. American Naturalist
115:718—742.

Glenn, T. C., W. Stephan, and M. J. Braun. 1999. Effects of a population bottleneck on
Whooping Crane mitochondrial DNA variation. Conservation Biology 13: 1097-1 107.

Hackney, EB. and J. B. McGraw. 2001. Experimental demonstration of an Allee effect in
American ginseng. Conservation Biology 15:129-136.

Hairston, N. G., Jr. 1996. Zooplankton egg banks as biotic reservoirs in changing
environments. Limnology and Oceanography 41:1087-1092.

Hairston, N. G., and B. T. J. De Stasio. 1988. Rate of evolution slowed by a dormant
propagule pool. Nature 336:239-242.

Hall, S. R., M. A. Leibold, D. A. Lytle, and V. H. Smith. 2004. Stoichiometry and
planktonic grazer composition over gradients of light, nutrients, and predation risk.
Ecology 85:2291-2301.

Hart, R. C., and I. A. McLaren. 1978. Temperature acclimation and other influences on
embryonic duration in the copepod Pseudocalanus sp. Marine Biology 45: 23-30

Hirst, A. G., and T. Kiorboe. Mortality of marine planktonic copepods: global rates and
patterns. Marine Ecology-Progress Series 230: 195-209

Hopper, K. R. and R. T. Roush. 1993. Mate finding, dispersal, number released and the
success of biological control introductions. Ecological Entomology 182321-331.

Jenkins, D. G., and A. L. Buikema. 1998. Do similar communities develop in similar
sites? A test with zooplankton structure and function. Ecological Monographs 68:421-
443.

J ersabek, C. D., and R. Schabetsberger. 1995. Resting egg production and oviducal
cycling in two sympatric species of alpine diaptomids (Copepoda: Calanoida) in
relation to temperature and food availability. Journal of Plankton Research 17: 2049-
2078

94

Katona, S. K. 1973. Evidence for sex pheromones in planktonic copepods. Limnology
and Oceanography 182574-583.

Keams, C. M., N. G. Hairston, and D. H. Kesler. 1996. Particle transport by benthic
invertebrates: Its role in egg bank dynamics. Hydrobiologia 332263-70.

Kent, A., C. P. Doncaster, and T. Sluckin. 2003. Consequences for predators of rescue
and Allee effects on prey. Ecological Modelling 162:233-245.

Kiorboe, T. 2006. Sex, sex-ratios, and the dynamics of pelagic copepod populations.
0ecologia 1482 40-50

Kiorboe, T. and E. Bageien. 2005. Motility patterns and mate encounter rates in
planktonic copepods. Limnology and Oceanography 50:1999-2007.

Kiorboe, T., E. Bageien, and U. H. Thygesen. 2005. Blind dating—mate finding in
pelagic copepods. II. The pheromone cloud of Pseudocalanus elongates. Marine
Ecology-Progress Series 300: 117-128.

Knapp, R. A., J. A. Garton, and O. Samelle. 2001a. The use of egg shells to infer the
historical presence of copepods in alpine lakes. Journal of Paleolimnology 25:539-
543.

Knapp, R. A., K. R. Matthews, and O. Samelle. 2001b. Resistance and resilience of
alpine lake fauna to fish introductions. Ecological Monographs 71:401-421.

Kuussaari, M., I. Saccheri, M. Camara, and I. Hanski. 1998. Allee effect and population
dynamics in the Glanville fiitillary butterfly. Oikos 82:384-392.

Lamont, B. B., P. G. L. Klinkhamer, and E. T. F. Witkowski. 1993. Population
fragmentation may reduce fertility to zero in Banksia goodii: a demonstration of the
Allee effect. 0ecologia 942446-450.

Leberg, P. L. 2002 Estimating allelic richness: effects of sample size and bottlenecks.
Molecular Ecology 11:2445-2449.

Leising, A.W., and J. Yen. 1997. Spacing mechanisms within light-induced copepod
swarms. Marine Ecology—Progress Series 1552127-135

Le Page, S. L., R. A. Livermore, D. W. Cooper, and A. C. Taylor. 2000. Genetic analysis
of a documented population bottleneck: introduced Bennett's wallabies (Macropus
rufogriseus rufogriseus) in New Zealand. Molecular Ecology 9: 753-763.

Leung, B., J. M. Drake, and D. M. Lodge. 2004. Predicting invasions: propagule pressure
and the gravity of Allee effects. Ecology 8521651-1660.

95

Lewis, M. A., and P. Kareiva. 1993. Allee dynamics and the spread of invading
organisms. Theoretical Population Biology 43:141-158.

Liebhold, A. and J. Bascompte. 2003. The Allee effect, stochastic dynamics and the
eradication of alien species. Ecology Letters 62133-140.

Liermann, M., and R. Hilbom. 1997. Depensation in fish stocks: a hierarchic Bayesian
meta-analysis. Canadian Journal of Fisheries and Aquatic Sciences 54:1976-1984.

Martcheva, M. and B. M. Bolker. 2007. The impact of the allee effect in dispersal and
patch-occupancy age on the dynamics of metapopulations. Bulletin of Mathematical
Biology 69:135-156.

Maruyama, T., and P. A. F uerst. 1985. Population bottlenecks and nonequilibrium
models in population genetics. 11. Number of alleles in a small population that was
formed by a recent bottleneck. Genetics 1112675-689.

Mayr, E. (1963). Animal species and evolution. Cambridge, Mass, Harvard University
Press.

McNaught, A. S., D. W. Schindler, B. R. Parker, A. J. Paul, R. S. Anderson, D. B.
Donald, and M. Agbeti. 1999. Restoration of the food web of an alpine lake following
fish stocking. Limnology and Oceanography 44:127-136.

Mittelbach, G. G., E. A. Garcia, and Y. Taniguchi. 2006. Fish reintroductions reveal
smooth transitions between lake community states. Ecology 872312-318.

Morris, D.W. 2002. Measuring the Allee effect: Positive density dependence in small
mammals Ecology 83214-20

Mosteller, F. and J. W. Tukey. 1977. Data analysis and regression: a second course in
statistics. Addison-Wesley, Reading, Mass.

Myers, R. A., N. J. Barrowman, J. A. Hutchings, and A. A. Rosenberg. 1995. Population
dynamics of exploited fish stocks at low population levels. Science 269:1106-1108.

Nei, M. 1987. Molecular evolutionary genetics. New York, Columbia University Press.

Nei, M., T. Maruyama, and R. Chakraborty. 1975. The bottleneck effect and genetic
variability in populations. Evolution 292 1-10.

Nihongi, A., S. B. Lovem, and J. R. Strickler. 2004. Mate-searching behaviors in the
freshwater calanoid copepod Leptodiaptomus ashlandi. Journal of Marine Systems
49265-74

96

Noel, H. L., S. P. Hopkin, T. H. Hutchinson, T. D. Williams, and R. M. Sibly. 2006.
Towards a population ecology of stressed environments: the effects of zinc on the
springtail F olsomia candida. Journal of Applied Ecology 43:325-332.

O'Brien, S. J. (1994). "A role for molecular genetics in biological conservation."
Proceedings of the National Academy of Sciences 91: 5748-5755.

Odum, P. E. 1959. Fundamentals of Ecology. Saunders, Philadelphia.

Pannell, J. R. and B. Charlesworth. 1999. Neutral genetic diversity in a metapopulation
with recurrent local extinction and recolonization. Evolution 532664-676.

Park, T. 1933. Studies in population physiology 11. Factors regulating initial grth of
T ribolium confusum populations. Journal of Experimental Zoology 65:17-42.

Parker, B. R., D. W. Schindler, D. B. Donald, R. S. Anderson. 2001. The effects of
stocking and removal of a normative salmonid on the plankton of an alpine lake.
Ecosystems 42334-345

Parker, B. R., F. M. Wilhelm, and D. W. Schindler. 1996. Recovery of
Hesperodiaptomus arcticus populations from diapausing eggs following elimination

by stocked salmonids. Canadian Journal of Zoology-Revue Canadienne de Zoologie
74:1292-1297.

Pennak, R.W. 1978. Freshwater invertebrates of the United States. Wiley, New York.

Podolsky, R. D. and R. B. Emlet. 1993. Separating the effects of temperature and
viscosity on swimming and water-movement by sand dollar larvae (Dendraster-
Excentricus). Journal of Experimental Biology. 1762 207-221.

Ramstad, K. M., C. A. Woody, G. K. Sage, and F. W. Allendorf. 2004. Founding events
influence genetic population structure of sockeye salmon (Oncorhynchus nerka) in
Lake Clark, Alaska. Molecular Ecology 13:277-290.

Robertson, A., Gehrs, C. W., Hardin, B. D., Hunt, G. W. 1974. Culturing and ecology of
Diaptomus clavipes and Cyclops vernalis. US. Environmental Protection Agency,
Washington DC. (EPA=660/3-74-006).

Robichaux, R. H., E. A. Friar, and D. W. Mount. 1997. Molecular genetic consequences
of a population bottleneck associated with reintroduction of the Mauna Kea
silversword (Argyroxiphium sandwicense ssp. sandwicense [Asteraceae]).
Conservation Biology 11:1 140-1 146.

Sakuratani, Y., K. Nakao, N. Aoki, and T. Sugimoto. 2001. Effect of population density

of Cylasformicarius (F abricius) (Coleoptera : Brentidae) on the progeny populations.
Applied Entomology and Zoology 36219-23.

97

Samelle, O., and R. A. Knapp. 2004. Zooplankton recovery after fish removal:
Limitations of the egg bank. Limnology and Oceanography 49: 1382-1392.

Serrano, D., D. Oro, U. Esperanza, and J. L. Tella. 2005. Colony size selection
determines adult survival and dispersal preferences: Allee effects in a colonial bird.
American Naturalist 1662E22-E31.

SPSS Inc. 1998. SYSTAT version 9.

Stemberger, R. S. 1995. Pleistocene refuge areas and postglacial dispersal of copepods of
the northeastern United States. Canadian Journal of Fisheries and Aquatic Sciences
52:2197-2210

Stephens, P. A., and W. J. Sutherland. 1999. Consequences of the Allee effect for
behaviour, ecology and conservation. Trends in Ecology and Evolution 142401-405.

Stoner, A. W. and M. Ray-Culp. 2000. Evidence for Allee effects in an over-harvested
marine gastropod: density-dependent mating and egg production. Marine Ecology-
Progress Series 202:297-302.

Strickler, J. R. 1998. Observing free-swimming copepods mating. Philosophical
Transactions of the Royal Society of London Series B—Biological Sciences 353: 671-
680

Strickler, J. R. and Hwang, J .-S. 1998. Matched spatial filters in long working distance
microscopy of phase objects. In P. C. Cheng, P. P. Hwang, J. L. Wu, G. Wang and H.
Kim [eds], Focus on multidimensional microscopy. World Scientific.

Taylor, C. M. and A. Hastings. 2005. Allee effects in biological invasions. Ecology
Letters 82895-908.

Templeton, A. R. (1980). "Modes of speciation and inferences based on genetic
distances." Evolution 34: 719-729.

Torke, B. 2001. The distribution of calanoid copepods in the plankton of Wisconsin
Lakes. Hydrobiologia 453:351-365.

Tsuda, A., and C. B. Miller. 1998. Mate-finding behaviour in Calanus marshallae Frost.
Philosophical Transactions of the Royal Society B: Biological Sciences 3532713—720.

VanLeeuwen, H. C. and E. J. Maly. 1983. Changes in swimming behavior of male

Diaptomus leptopus (Copepoda, Calanioda) in response to gravid females. Limnology
and Oceanography 36: 1188-1195.

98

Veit, R. R. and M. A. Lewis. 1996. Dispersal, population growth, and the Allee effect:
dynamics of the house finch invasion of eastern North America. The American
Naturalist 148: 255-274

Wade, M. J ., and D. E. McCauley. 1988. Extinction and recolonization: Their effects on
the genetic differentiation of local populations. Evolution 422995-1005.

Watras, C. J. 1983. Mate location by diaptomid copepods. Journal of Plankton Research
52417-425.

Watras, C. J. and J. F. Haney. 1980. Oscillations in the reproductive condition of
Diaptomus leptopus (Copepoda: Calanoida) and their relation to rates of egg-clutch
production. 0ecologia 45294-103.

Williamson, C. E., and N. M. Butler. 1987. Temperature, food and mate limitation of
copepod reproductive rates: separating the effects of multiple hypotheses. Journal of
Plankton Research 9:821-836.

Wright, S. 1931. Evolution in mendelian populations. Genetics 16197-159.

Wright, S. 1938. Size of population and breeding structure in relation to evolution.
Science 872430-431.

Wright, S. 1943. Isolation by distance. Genetics 28:114-138.
Yen, J ., M. J. Weissburg, and M. H. Doall. 1998. The fluid physics of signal perception

by mate-tracking copepods. Philosophical Transactions of the Royal Society of
London Series B—Biological Sciences 353: 787-804

99

    
  

       

VR

iiiiiliiijigiiiiiiili

tilt

 

95