Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Test of alternative domains of attraction in the dynamics of a fishless oligotrophic lake Ouimet, Chantal 1998-06-24

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
831-ubc_1998-346056.pdf [ 10.26MB ]
Metadata
JSON: 831-1.0074832.json
JSON-LD: 831-1.0074832-ld.json
RDF/XML (Pretty): 831-1.0074832-rdf.xml
RDF/JSON: 831-1.0074832-rdf.json
Turtle: 831-1.0074832-turtle.txt
N-Triples: 831-1.0074832-rdf-ntriples.txt
Original Record: 831-1.0074832-source.json
Full Text
831-1.0074832-fulltext.txt
Citation
831-1.0074832.ris

Full Text

Test of alternative domains of attraction in the dynamics of a fishless oligotrophic lake By Chantal Ouimet B. Sc., Universite de Montreal, 1982 M. Sc., Universite de Montreal, 1986 A-THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Department of Zoology) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA July 19,1998 © Chantal Ouimet, 1998 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, 1 agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of ~Z.Co[o The University of British Columbia Vancouver, Canada DE-6 (2/88) ABSTRACT The theory of domains of attraction (alternative "stable" states) allows variability, thresholds and persistence as integral parts of ecological system functioning. This thesis investigates the potential for alternative domains of attraction in a zooplankton community. Two alternative states have been proposed for the zooplankton community of small Ashless, oligotrophic, mountain lakes. In the "competition state", Daphnia (Cladocera) outcompete solitary rotifers and impede predator recruitment. This state is persistent and resilient to disturbance. In the "predation state", predation by Chaoborus (Diptera) on both prey types alters community dynamics to favor rotifers over Daphnia. Persistence of the predation state requires reduced prey competition, enhanced young predator recruitment in summer and predator survival in high densities overwinter. I carried out graded field experiments using reduced Daphnia densities and predator additions in the spring to generate and test the predator state. I monitored lake and enclosure communities for three consecutive years using an adaptive sampling concept. Chaoborus recruitment was enhanced in enclosures and the new predator cohorts survived overwinter in high densities. However, developmental delays prior to the winter period prevented persistence of the predator state over several generations. Enhanced predator densities in the spring, as well as cold temperature, delayed Daphnia population onset and increase, which released solitary rotifer populations needed to feed young ii predator recruits in early summer. However, in the higher predator treatments, both prey types stayed depleted till late summer which resulted in delayed predator development. Laboratory experiments showed that although Chaoborus americanus can pupate at 5°C, they could not metamorphose into adults below 9°C. Cold water can delay their reproductive phase and delay timing of young predator recruitment. I conclude that Shirley Lake, under current nutrient levels, does not have two domains of attraction. Nonetheless the presence of a threshold between states enlarges the window for coexistence of weaker competitors or rare species. Thresholds lead to alternative domains of attraction in some systems, and to transient state in others. From a management perspective, extended transient states can either lead to misleading interpretation and erroneous interventions if permanent changes are expected or be used as tools to produce temporary changes. iii Table of Contents ABSTRACT ii TABLE OF CONTENTS iv LIST OF TABLES viii LIST OF FIGURES x ACKNOWLEDGMENTS xiii CHAPTER 1 1 GENERAL INTRODUCTION 1 1.1 DOMAINS OF ATTRACTION AND THRESHOLDS IN THEORY AND IN THE FIELD 1 1.2 ZOOPLANKTON COMMUNITY COMPOSITION AND FUNCTIONING: DAPHNIA DOMAIN VERSUS CHAOBORUS DOMAIN 13 CHAPTER 2 30 TRANSIENT STATE OR DOMAIN OF ATTRACTION: TESTING THE PERSISTENCE OF THE CHAOBORUS STATE2.1 INTRODUCTION 30 2.2 MATERIAL AND METHODS 32 2.2.1 Field site and enclosure design 3 2.2.2 Experimental design 9 2.2.3 Variable sampling interval: towards an adaptive sampling design 45 2.2.4 Sampling methods, identification, and counts for Chaoborus: 48 2.2.5 Laboratory experiments 50 2.2.6 Initial experimental conditions and general seasonal patterns in enclosures 51 2.2.7 Predictions based on experimental design 56 iv 2.3 RESULTS 61 2.3.1 Second instar predator recruitment in summer 1992: density, duration 61 2.3.2 Fourth instar ability to resist starvation: survival in the laboratory and in field experiments 7 2.3.2.1 Survival in laboratory experiments at 5°C 62.3.2.2 Survival overwinter in field enclosure experiments 8 2.3.3 A signal in transition: third instar and fall fourth instar predator dynamics in 1992 73 2.3.4 Result summary 82 2.4 DISCUSSION 4 2.5 CONCLUSION 9 CHAPTER 3 91 PREY DYNAMICS, SHORT TIME SCALES, AND PREDATOR RECRUITMENT 91 3.1 INTRODUCTION3.2 MATERIALS AND METHODS 94 3.2.1 Field experiments3.2.2 Identification and counts 6 3.2.3 Predictions for prey dynamics based on experimental design 97 3.3 RESULTS 101 3.3.1 General trends in prey population dynamics 103.3.2 Daphnia population dynamics: influence of spring predator density and temperature ...107 3.3.2.1 Impact of the 1992 spring predator density gradient on Daphnia densities 103.3.2.2 Daphnia population onset: influence of water temperature and predator density in the spring 113 3.3.3 Relationship between Daphnia population and solitary rotifer population dynamics...121 3.4 DISCUSSION 133.4.2 Rotifer population dynamics 136 3.5 CONCLUSION 138 V CHAPTER 4 140 CHAOBORUS PUPATION IN COLD WATER: IMPLICATIONS FOR LIFE HISTORY, DISTRIBUTION AND POPULATION DYNAMICS 140 4.1 INTRODUCTION 144.2 MATERIAL AND METHODS 141 4.2.1 Field collection and laboratory set up 144.2.2 Data analysis methods 4 4.3 RESULTS 145 4.4 DISCUSSION 150 4.5 CONCLUSIONCHAPTER 5 156 GENERAL DISCUSSION AND CONCLUSION 155.1 The makings of an extended transient state 6 5.2 Dynamical thresholds: the role of nutrient availability, temperature and species composition ; 161 5.3 Alternative domains and states: importance of the threshold perspective 165 5.4 Hysteresis: one threshold when going up, another when going down 167 5.5 Management issues in a threshold perspective 169 CONCLUSION 171 APPENDDC A Enclosure construction design 173 APPENDIX B Method for enclosure fill up with pumps 174 APPENDIX C Comments on predator time series 1992-1994 175 APPENDIX C-l Monthly sample time series for total Chaoborus first instar larva density in Shirley Lake and in experimental enclosures from 1992 to 1994 177 APPENDIX C-2 Monthly sample time series for total Chaoborus second instar larva density in Shirley Lake and in experimental enclosures from 1992 to 1994 178 APPENDIX C-3 Monthly sample time series for total Chaoborus third instar larva density in Shirley Lake and in experimental enclosures from 1992 to 1994 179 APPENDIX C-4 Monthly sample time series for total Chaoborus fourth instar larva density in Shirley Lake and in experimental enclosures from 1992 to 1994 180 vi APPENDIX D Monthly sample time series for total Daphnia density in Shirley lake and in experimental predator addition enclosures from 1992 to 1994 181 APPENDIX E Monthly sample time series for total solitary rotifer density in Shirley lake and in experimental predator addition enclosures from 1992 to 1994 182 BIBLIOGRAPHY 183 vii LIST OF TABLES Table 1.1 Requirements for the persistence of Chaoborus state as an alternative domain of attraction to the Daphnia domain 27 Table 2.1 Experimental design for predator and nutrient additions 42 Table 2.2 Timing of enclosure recruitment failure 57 Table 2.3 Predictions for the relationships in predator densities and in prey densities between the lake and the enclosures, and between the low and high treatments in relation to the experimentally imposed predator gradient 59 Table 2.4 Test of the hypothesis that enclosures have crossed the threshold and switched to Chaoborus state: Chaoborus second instar larvae summer recruitment period 1992 64 Table 2.5 Results from overwinter survival laboratory experiments on Chaoborus 69 Table 2.6 Test of the hypothesis that enclosures have crossed the threshold and switched to Chaoborus state: Chaoborus fourth instar larvae mean density, April-May 1993 72 Table 2.7 Test of the hypothesis that enclosures have crossed the threshold and switched to Chaoborus state: Chaoborus third instar larvae, summer recruitment period 1992 78 Table 2.8 Test of the hypothesis that enclosures have crossed the threshold and switched to Chaoborus state: Chaoborus fourth instar larvae, fall 1992 80 Table 2.9 Overall results of the impact of the spring fourth instar predator density gradient on the predator dynamics throughout the life cycle83 Table 3.1 Predictions for the relationships in prey densities between the lake and the enclosures, and between the low and high treatments in relation to the experimentally-imposed predator gradient 99 viii Table 3.2 Impact of fourth instar predator density gradient on Daphnia population increase: Delays (in weeks) in enclosure Daphnia population in reaching densities similar to those found in the lake at the time when first instars appeared 110 Table 3.3 Test of the hypothesis that enclosures have crossed the threshold and switched to Chaoborus state: Testing the difference in Daphnia density (June 1992) between the lake and the enclosures, and between the low and high predation enclosures 112 Table 3.4 Test of the hypothesis that enclosures have crossed the threshold and switched to Chaoborus state: Testing the difference in Daphnia density (August 1992) between the lake and the enclosures, and between the low and high predation enclosures 120 Table 3.5 Test of the hypothesis that enclosures have crossed the threshold and switched to Chaoborus state: Testing the difference in solitary rotifer density between the lake and the enclosures, and between the low and high predation enclosures, in relation to Daphnia densitiesl24 Table 3.6 Overall results of the impact of the 1992 spring density gradient in fourth instar predators on the prey dynamics 129 Table 4.1 Laboratory experimental conditions for raising Chaoborus 142 Table 4.2 Status of Chaoborus larvae in laboratory experiments at different temperatures 146 Table 4.3 Duration of pupation at the individual level 147 ix LIST OF FIGURES Figure 1.1 Schematic representation of global stability, local stability and domains of attraction 3 Figure 1.2 Small prey: solitary rotifers (Rotifera) 14 Figure 1.3 Large prey: Daphnia rosea (Cladocera: Daphnidae) 16 Figure 1.4 Relative size: Daphnia(on the right) versus solitary rotifer Keratella (on the lower left) 18 Figure 1.5 The predator Chaoborus sp. (Diptera: Chaoboridae): head and jaw close up 19 Figure 1.6 Family portrait: Chaoborus larval instars and pupa 20 Figure 1.7 Schematic of interactions in the Daphnia domain of attraction: the competition state 23 Figure 1.8 Schematic of interactions in the hypothesized Chaoborus domain of attraction: the predation state 24 Figure 2.1 Map of Shirley Lake in the Malcolm Knapp UBC Research Forest, Maple Ridge, B.C., Canada (after Butler, 1990) 34 Figure 2.2 Life histories of Chaoborus americanus and C. trivittatus in Shirley Lake from spring 1992 to spring 1995 36 Figure 2.3 Enclosure set up and experimental design 38 Figure 2.4 Daphnia reduction: densities of Daphnia in enclosures after water pumping and before predator additions. Density in lake for the same time period is provided for reference 41 Figure 2.5 Initial densities of fourth instar predators in Shirley Lake and in the enclosures on May 26th, 1992 52 Figure 2.6 Large prey density reduction: percentage of Daphnia density removed from enclosures relative to Daphnia density in the lake after predator additions (May 26th, 1992) 54 Figure 2.7 Duration of Chaoborus populations in experimental enclosures and in Shirley lake 55 x Figure 2.8 Figure 2.9 Figure 2.10 Figure 2.11 Figure 2.12 Figure 2.13 Figure 2.14 Figure 3.1 Figure 3.2 Figure 3.3 Figure 3.4 Figure 3.5 Figure 3.6 Recruitment of second instar predators in summer 1992 in relation to the density gradient in fourth instar predators in spring 1992 62 Duration of the recruitment period for second instar predators in 1992 66 Chaoborus fourth instar densities in spring 1993 in relation to the experimental density gradient in fourth instar predators in spring 1992 70 Proportion of second instars surviving to fourth instars (summer 1992 to spring 1993) in relation to the density gradient in fourth instar predators in spring 1992 74 Signal transmission through the life history of the larval instars of Chaoborus americanus. The x-axis (except for panel (a)) represents the experimental density gradient in fourth instar predators in spring 1992 76 Duration of the main recruitment period for third instar larvae of C. americanus in Shirley Lake and in enclosures 79 Relative difference in fourth instar Chaoborus density between fall 1992 and the subsequent spring (1993) 81 Daphnia density on the date when first instar predators appeared in Shirley Lake and in the enclosures each year in relation to the density of fourth instar predators in each spring 102 Density of solitary rotifers on the date when first instar predators appeared in Shirley Lake and in the enclosures each year in relation to the density of fourth instar predators each spring 103 Time series for total Daphnia density May to October 1992 following a variable sampling interval 104 Time series for total Solitary rotifer density May to October 1992 following a variable sampling interval 105 Initial Daphnia densities in the lake and in the enclosures after predator additions (May 26th, 1992) 108 Mean density of Daphnia (for Julian days 154 to 168, June 1992) in relation to the experimental gradient in fourth instar predator in spring 1992 111 xi Figure 3.7 Figure 3.8 Figure 3.9 Figure 3.10 Figure 3.11 Figure 3.12 Figure 3.13 Figure 4.1 Figure 4.2 Figure 5.1 Figure 5.2 Delay in Daphnia population onset in 1992 to 1994 in relation to (A) mean water temperature in Shirley Lake in springtime (mid-May to mid-June), (B) predator density in Shirley Lake in spring 1992 to 1994 115 Maximum and yearly mean Daphnia density in 1992 to 1994 in relation to yearly mean temperature 116 Delay in Daphnia onset in the spring in relation to the density of fourth instar predators in springtime of each year 118 Daphnia densities in August 1992, when third instar predators reached their maximum density, in relation to the density gradient in fourth instar predator in spring 1992 119 Mean density of solitary rotifers in relation to Daphnia mean densities for Julian days 154 to 168 (June 1992) 123 Solitary rotifer densities and Daphnia densities in 1992, initially in May, during young predator recruitment in June, and during third instar recruitment in August in relation to density gradient in fourth instar predators in spring 1992 126 Solitary rotifer densities in relation to Daphnia maximum densities after enclosures have lost their predator cohort (A) for enclosures that failed in 1993, (B) for enclosures that failed in 1994 128 Comparison of survival rates for Chaoborus americanus larvae raised at 5, 9, and 12 °C 148 Schematic representing the effect of developmental rate acceleration due to fluctuating temperature in the low part of the range of developmental temperature of an insect 153 Schematic representation of the bottlenecks in Chaoborus recruitment 158 Schematic representation of a hysteresis loop and the area of dynamic unpredictability in system with two domains of attraction.168 xii ACKNOWLEDGMENTS This project is the result of the involvement of numerous people. I am deeply indebted to Dr. Bill Neill who shared openly his knowledge about science and more importantly how to do science. I would like to express my gratitude for his support and his anchoring influence throughout this long adventure. I appreciated his generosity in allowing me extensive use of his laboratory and office space, and of his field and laboratory equipment. I owe a thousand thanks to Danusia Dolecki, for her friendship, for her great technical abilities, and for sharing her vast knowledge of all that is aquatic; without her, a green horn such as I would have been swept under by the complexity of the task. I acknowledge with gratitude the help provided by my research committee: Dr. Don Ludwig, Dr. Bill Neill, Dr. Tom Northcote, Dr. John Post, Dr. John Richardson, Dr. Dolph Schluter, and Dr. Carl Walters. They encouraged me to focus and clarify my ideas. The comments of Drs. Ludwig, Neill, Richardson and Schluter greatly improved my writing and the resultant thesis while allowing me the freedom to make my own mistakes. I am thankful for the diligence with which they perused successive versions of this work. I thank Carolyn Cornish, Bill Graham, Rose Murphy, Ole Olson, and Peter Troffe, student employees who participated enthusiastically in the field and laboratory work. Many thanks to Dr. Ken Hall, Dr. Paul. J. Harrisson, Dr. Al Lewis, Dr. Don McPhail, Dr. Bill Neill, Dr. Tom Northcote, Dr. Tim Parsson, Dr. John Post who generously lent me field and laboratory equipment. Thanks to Catriona, John and Quedong for help on nutrient analysis, and to Keith McDougall for microscopic photography and image handling. I want to thank staff from the Zoology Department, Fisheries Centre and Malcolm Knapp Research Forest who guided me through the administrative jungle or helped with some technical aspects of the project. I must acknowledge Dr. Carl Walters, my supervisor, for the laboratory space he provided and for generously funding this project out of his research grants. Other funding for this project was provided by N.S.E.R.C. through a two-year scholarship, by the University of British Columbia through a two-year University Graduate Fellowship and several Teaching Assistantships. Supplemental support was provided as student loans by the Government of Canada and the Government of Alberta. Finally, I offer my sincere thanks to over three dozen volunteers who visited Shirley Lake over the years and helped with installing, sampling (especially at night), and removing the enclosures: Rob Ahrens, Lance Barrett-Lennard, Shannon Bennet, Sarah Beukema, Alice Cassidy, Maggie Cobbett, Carolyn Cornish, Danusia Dolecki, Lech Dolecki, Richard Dolecki, Reuven Dukas, Melissa Fletcher, David Ghan, Kathy Gorkoff, Elvira Harms, Steve Heard, David Hik, Wes Hochachka, Leonardo Huato, Xavier Lambin, Anna Lindholm, Karl Mallory, Keith McDougall, Maura Mclnnis, Jeff and Irene Mclnnon, Matt McLeod, Bill Neill, Jean-Michel Pianotti, John Post, John Pritchard, Jordan Rosenfeld, Dick Repasky, Beth Scott, Susan Shirley, Lisa Thompson, Carl Walters, David Ward, Crystal, Forrest, Kerry, Maria, Michael, Mike, Regina, William. To those I omitted by oversight, I thank you earnestly for your help and support. Thank you all. xiii TO KEITH, Nothing is more patient than love. A LYSE, CLAUDE, ALAIN, LYNE, MICHEL, MANON, Oh! combien de loin Mais avec grand soin Vos pensees ont voyagees Et m'ont accompagnees Tout au long de mon sejour, De mes nombreux detours, Dans les montagnes du vecu, Au pays de l'inconnu. xiv "The history of ecology is a history of changing criteria for imposing order on nature and resisting the alternative that all is really chaotic and contingent" Sharon KINGSLAND (1985) "Modeling nature: Episodes in the history of population ecology, p.5" xv CHAPTER 1 GENERAL INTRODUCTION Ecological systems persist in the face of repeated disturbances despite variation in species abundance and composition. However, emphasis on linear dynamics and trends in interpretation of ecological data fails to take account of such variation. The linear model represents well systems with continuous behavior and for which a small disturbance always results in a small response of the system. In nonlinear dynamics, response of the system is not always proportional to the size of the disturbance, but the model is still continuous. Recently, some ecological systems have been shown to have discontinuous behaviors or thresholds where even a small disturbance can generate large changes in dynamics, behaviors which cannot be explained in terms of linear dynamics. 1.1 DOMAINS OF ATTRACTION AND THRESHOLDS IN THEORY AND IN THE FIELD The concept of multiple domains of attraction can integrate both variability and the potential for discontinuities in ecological dynamics. Domains are represented in phase space, a graph where dynamical trajectories resulting from the interactions of two variables of the system can be represented. Examples of phase space are predator-prey graphs or competitor interaction graphs with their 1 dynamical isoclines and their equilibrium points. A domain of attraction is an area of phase space that includes the set of all trajectories leading back to the attractor (or reference state, or dynamic) after a disturbance (Figure 1.1). A system with multiple domains of attraction has local, but not global stability. Global stability occurs in a system with a single stable equilibrium (Figure 1.1-A); all trajectories in phase space lead to this attractor. On the other hand, local stability means that more than one stable equilibrium is present (Figure 1.1-B); different trajectories lead to different attractors, and thus the system can have more than one behavior. Domains are separated by boundaries, which are thresholds for the variables of interest. Ecological systems have been modeled using the concept of multiple domains of attraction to describe community dynamics and explain discontinuities (e.g., Ludwig et al. 1978; Peterman et al. 1979; Harrison 1986; Adams and DeAngelis 1987; Scheffer 1989; 1991a; Carpenter and Cottingham 1997; 1998). However, the evidence for multiple domains of attraction in the field is sparse (Edmonson and Lehman 1981; Sinclair 1995; Scheffer 1998), and questions arise: how ubiquitous are they and do they play an important role in ecological systems? Thresholds play an important role at several scales in ecology. Furthermore, thresholds at one scale can affect dynamics at the scale above. At the physiological level (individual level), ectotherms, such as insects or zooplankton, have temperature thresholds. The thresholds are associated with on/off switches, for example, a temperature above which development can take 2 Figure 1.1 Schematic representation of global stability, local stability and domains of attraction. A) Global stability: all trajectories lead to a single attractor Predator density Prey density B) Local stability: different trajectories lead to different attractors. The bold line represents the boundary separating the two domains of attraction. Predator density Prey density 3 place and below which there is no development (Hoffmann 1985). Other thresholds are associated with additive processes (time or energy dependent), such as accumulation of degree-days or accumulation of body mass, where development takes place but is only completed when the time or energy threshold is crossed. When the degree-day threshold is reached, for example, zooplankton can molt to move between stages, and insects can metamorphose into adults and reproduce. These physiological thresholds often lead to synchronized emergence of aquatic insects that facilitate mating and reduce predation losses. Furthermore, from a predator's perspective, synchronized emergence improves predation success and potentially reproductive success. Thus, physiological thresholds have implications for population dynamics. Moreover, they can have a rippling effect on other species in the system and potentially change community dynamics. How common are discontinuous behaviors at the community scale? Ecological systems were modeled as multiple domains of attraction starting in the mid-1970s (Pasture productivity: Noy-Meir 1975; competition in birds: Gilpin and Case 1976; Budworm epidemics: Ludwig et al. 1978; Fish and insect: Peterman et al. 1979). The early model by Noy-Meir (1975) showed alternative domains of attraction, which he called "alternative stable states". He suggested that different herbivore characteristics (e.g. sheep versus goats as grazers) would lead to different results, with efficient grazers more likely to create discontinuity in the behavior of the system. In such a system, overgrazing could reduce the 4 productivity of a pasture, but removing grazing pressure would not necessarily return productivity to higher levels (hysteresis). He also suggested management practices which would preserve pasture productivity. Ludwig et al. (1978), emphasized the importance of the interactions of different scales, i.e., fast and slow variables, in their simulation model of spruce budworm outbreaks. The build-up of the crown foliage could take several decades (slow variable) before it would reach a level at which the spruce budworm larvae could escape the predation influence and produce a sudden population outbreak (fast variable) which would disappear within a single decade. With this model, the authors suggested that spraying the insect larvae would maintain the forest at high food quality, thus maintaining the epidemic conditions year after year. Other models have been developed more recently (Fish populations: Adams and DeAngelis 1987; pike-bream-phytoplankton-macrophytes: Scheffer 1989; Jeppesen et al. 1990 in Blindow et al. 1993; phytoplankton-macrophytes: Moss 1990; 1990; 1991a; Blindow et al. 1993; Fish-zooplankton: Rudstam et al. 1993; lake eutrophy-dystrophy: Carpenter and Cottingham 1997; 1998). Each of these models is based on phenomena taking place over more than one scale, temporal or spatial. For example, in the phytoplankton-macrophyte system, phytoplankton has several generations per year (fast variable), while macrophytes have only one (slow variable). The fast variable (phytoplankton) under a small disturbance (increased phosphorus loading) can feed back and 5 multiply the effect of the small disturbance into a large observable impact on the species, population or community under observation. The phytoplankton bloom can eliminate macrophytes through shading, a competitive interaction. The impact of a disturbance on a system depends both on characteristics of the species in presence and on the size and characteristics of the disturbance itself, characteristics such as type (pulse, press), or extent (e.g. time duration, space coverage, density range) (Pahl-Wostl 1995; Grimm and Wissel 1997). Thresholds have been observed in the field and described more frequently over the last decade. Two systems have been studied in sufficient details to define domains of attraction: 1) phytoplankton vs. macrophyte phases in shallow lakes (Moss 1990; Scheffer 1998), and 2) woodland vs. savannas in Serengeti (Africa) (Dublin et al. 1990; Sinclair 1995). Other systems present the mechanisms leading to potential or demonstrated thresholds in goose-plant interactions (Hik et al. 1992), piscivore-planktivore fish communities (Persson and Greenberg 1990; Mittelbach et al. 1995; Olson et al. 1995), and in a zooplankton community (Neill 1988a; 1988b). I review below some of these examples to portrait the type of interactions and the amount of information required to evaluate the dynamics of systems with discontinuities. A well studied system leading to two domains of attraction occurs in shallow lakes where dominance alternates between phytoplankton and macrophyte phases, between turbid and clear water phases. Enclosures and whole lake 6 experiments have revealed a threshold between the turbid state, due to large concentrations of phytoplankton, versus the clear water state, where macrophytes dominate (Scheffer 1989, 1990,1991a, 1998; Moss 1990; Blindow et al. 1993). Relative abilities in plant nutrient uptake and the effect of zooplankton grazing affect the state of the lake. When nutrient loading increases sufficiently, phytoplankton production can overcome losses to zooplankton grazing. Large blooms of phytoplankton decrease water clarity, thus shading the macrophytes, which die. When macrophytes disappear, zooplankton lose their refugia from fish predation, and the turbid state can become persistent. In the plant-fire-elephant interactions which lead to alternative states of savannas and woodlands in the Serengeti landscape, social history (rinderpest epidemic, ivory and slave trades, poaching) and natural history (wildebeest migration, grazing interaction and impact on plant community) were interlaced to gain insights on past landscapes, and give insight for current management for the preservation of this complex and dynamic ecosystem (Dublin et al. 1990; Dublin 1995; Sinclair 1995). Ecological studies revealed that fire, often set by human activities, could increase mature tree mortality while elephant grazing increased tree sapling mortality. When fires are frequent, savannas are created. When fires are rare, savannas can be maintained by a large elephant population. Only when both elements decreased did savannas change to woodlands. In the 1890s, a rinderpest epizootic reduced savanna grazers and destroyed tribal cattle herds. The latter reduced human activity and fire frequency. At the same time, 7 slave and ivory trade further reduced human-set fires and reduced elephant populations. This combination led to increased cover of woodlands in the East African landscape. In the 1950s and 1960s, the human population increased dramatically, thus increasing fire frequency. Moreover, the increase in human populations also resulted in the compression of elephants to smaller areas. This led to a decline in woodland cover. In the Serengeti-Mara ecosystem, the rebuilding of large populations of wildebeests after the rinderpest reduced dry tall grass fuel and reduced fire frequency but woodlands did not recover because of elephants grazing on tree saplings. Thus, two alternative states exist in the Serengeti landscape. In some systems, the information gathered so far reveals the existence of thresholds without determining if the states present are domains of attraction. Factors that differentially influence individual and population growth rates of the predator and prey species can affect community structure (Persson 1987; Persson 1988 : cited in Olson et al, 1995; Olson et al. 1995). In these studies different species of fish were observed, yet common characteristics arise. 1) The predator life history has an ontogenetic shift in feeding, i.e., smaller/younger individuals eat different resources than older ones. 2) Larger predators feed on prey species that compete for resources with the smaller-sized predators. 3) The predator ontogenetic shift is flexible, i.e., it is based not on a specific time or size but is rather based on relative size between the predator species individuals and the species which can function either as a prey or a competitor. 4) Young-of-the-8 year of predatory fish are planktivores until they reach a size at which they can feed on fish. 5) Young predators must compete with fish specialized in planktivory, which are better competitors when a food resource is limiting (Olson et al. 1995). This competition is detrimental to the young predators' growth rate, and their shift to piscivory is delayed, favoring a planktivore-dominated community. When plankton resources are not limiting, the young predators grow quickly to the relative size at which they can start feeding on the planktivorous fish. The community is then piscivore-dominated and zooplankton can bloom. Here again, temperature change is not necessary for a shift to take place, but could play a role, with a cool year reducing plankton productivity, thus increasing the likelihood of competition, and favoring a planktivore dominated community. Simple differences in consumption versus production rates, such as in the Cisco-Daphnia interaction in Lake Mendota (Rudstam et al. 1993), can also result in discontinuous community dynamics. In early spring, Cisco (Coregonus artedi), a cold-water planktivore fish, consume Daphnia at a higher rate than the rate Daphnia can reproduce in cold water. Thus, planktivory controls the prey population. As water temperature increases, Daphnia reproduction rate increases to levels exceeding the consumption rate of the planktivore. Thus, later in the spring, Daphnia population escapes control by Cisco, blooms and reduces phytoplankton populations, leading to the classic clear-water phase where water transparency increases. The rates of the predator and the prey are mediated by 9 both a biotic factor, density of planktivores, and an abiotic factor, water temperature. A switch in the dominance of one factor, predation on Daphnia, to another, recruitment of Daphnia, due to the physiological ecology of the two interacting species leads to changes in population and community dynamics. There are at least three reasons why few examples of thresholds in community dynamics are present in the literature: first, a perspective centered on stability combined with multiple definitions of stability, second, the need for ecologists to work at multiple scales and with multiple factors, third, the need for researchers and managers for a large amount of information and synthesis to link those different scales and factors. Holling (1973 , p.15), an early proponent of the first point, suggested that a very different view of the world could be obtained "if we concentrate on the boundaries of the domains of attraction rather than on equilibrium states." However, based on the deeply-rooted belief in the "balance of nature" (Egerton 1973), research has long been focused on the pursuit of ecological system equilibria, a view represented by the use of the expressions "multiple stables states" and "alternative stables states". A major assumption within this perspective is that ecological systems are able to reach and stay at equilibrium. Yet, ecological systems, under repeated disturbances both from biotic and abiotic variables, rarely achieve equilibrium. Over time, the "concept" of stability developed into a frenzy of definitions. While Lewontin (1969) recognized five definitions related to the notion of ecosystem stability, Grimm and Wissel (1997) catalogued 70 concepts and 163 definitions relating to stability. 10 They summarized this multitude into six main concepts: constancy, elasticity, persistence, resilience, resistance, and domain of attraction, of which the last four will be useful to this study. Persistence is a characteristic of an ecological system in which the system is maintained through time as an identifiable entity, without reference to its dynamics. Resilience is the ability of a system to return to the reference state after a temporary disturbance. Resistance is the ability of an ecological system to stay essentially unchanged despite the presence of disturbances. A domain of attraction is the set of conditions from which the reference state (or dynamic) can be reached again after a temporary disturbance. Ecological systems vary continually, but their variability is generally bounded. Determining those boundaries could be more informative than pursuing ever evasive equilibria. Secondly, the focus on a single factor, such as predation, or competition, or production, in isolation (as criticized by Polis 1994), and often over a single scale of observation (as criticized by Holling 1992; Polis 1994; Schneider 1994) is another reason for the fact that communities are rarely described in terms of multiple domains of attraction. For example, in the 1980s, in aquatic ecology, there was a controversy about which of top-down (predation) or bottom-up (production) was the dominant force in structuring communities. Trophic cascades imply a dominant impact of predation on the structure of communities (Fretwell 1987; Carpenter 1988a; Carpenter and Kitchell 1993), while nutrient levels and primary production imply bottom-up structuring factors. Similarly, 11 competition was seen as the driving force within guilds until Holt (1977) proposed the idea of apparent competition, where dynamics resembling a competitive interaction between two species of a guild are in fact driven by a predator switching prey species, based on the most abundant species present. In such a case predation, not competition, structures the community. A more fruitful approach looks at the relative strength of these processes under different conditions to determine what the community structure or dynamic will be (e.g. Neill and Peacock 1980; Leibold 1989; McQueen et al. 1989; Power 1992; Hunter 1992b). This approach implies the study of multiple processes simultaneously, which means more complex studies involving multiple species or categories, at more than one scale of observation, temporal or spatial (Addicott et al. 1987; Polis 1994; Polis and Strong 1996). Thirdly, the inclusion of multiple factors and multiple scales in the study of ecological systems increases the difficulty of the collection, representation and analysis of data (Addicott et al. 1987; Polis 1994). Large amounts of information regarding community interactions and the timing of events are required to detect and test the presence, resilience and persistence of alternative states. In turn, thoughtful synthesis of this information allows the assessment of two criteria that are essential to demonstrate the existence of multiple domains of attraction in a system. The first is to determine the presence of a threshold between two states, while the second is to demonstrate the persistence over several generations of each of the two (or more) states. Some evidence has been obtained 12 for two states in a zooplankton community composed of Chaoborus, the predator, and Daphnia and rotifers, the prey. This system is the focus of my research. 1.2 ZOOPLANKTON COMMUNITY COMPOSITION AND FUNCTIONING: DAPHNIA DOMAIN VERSUS CHAOBORUS DOMAIN In Gwendoline Lake, in the Malcolm Knapp Research Forest (British Columbia, Canada), Neill (1981b; 1985; 1988a) demonstrated the presence of a threshold which highlighted the potential of the zooplankton community to harbor two domains of attraction. The lake community was generally found in the competition domain, where competition between Daphnia and rotifers was the main interaction. Under experimental conditions in enclosures, the community was observed to switch to a different state, where predation of Chaoborus on both prey types became more important than the competition between prey. The switch was based on sufficient resources being directed to the rotifers in late spring. Rotifers (Figure 1.2) are generally suspension feeders. They feed on a range of small particles (0.5 - 20 um) (Bogdan and Gilbert 1982; Gilbert 1988a; Arndt 1993), which include bacteria, protista, small ciliates and small algae. Genera, such as Polyarthra, feed on larger particles (1-40 um) (Arndt et al. 1990; Gilbert and Jack 1993) and have poor ability to filter bacteria and small particles (Bogdan and Gilbert 1982). Under favorable conditions, rotifers reproduce parthenogenetically, 13 Figure 1.2 Small prey: solitary rotifers (Rotifera). ROTIFERS: general body size: 50 - 200 urn Suspension feeders: on 0.5- 20 um particle range Keratella sp (Brachionidae) with females quickly producing new females without sexual reproduction. They reproduce sexually when conditions deteriorate (Gilbert 1988a). In early spring, their population densities bloom before Daphnia hatch out of resting eggs (Gilbert 1988a). As the Daphnia population increases in late spring, the populations of rotifers decline (Neill 1985; Gilbert 1988a; 1988a). Rotifers bloom again in the fall when Daphnia enter the resting egg stage and the latter disappear from the water column (Gilbert 1988a). Daphnia are also suspension feeders. They can ingest a wide variety of particles — bacteria, ciliates, phytoplankton and even rotifers — of different sizes (Range: 0.5 - 60 um) (J0rgens0n 1966, cited in Hebert 1978; Jiirgens 1994a) with preference for cells 20 um or less (Edmonson 1959). Daphnia hatch from resting eggs in spring. In the Research Forest lakes, Daphnia rosea (Figure 1.3) generally appear in the samples in May, and bloom to high densities through parthenogenetic reproduction by the end of June (Walters et al. 1987). In the fall, males are produced, sexual reproduction takes place, and resting eggs are released to survive unfavorable conditions (Edmonson 1959) such as those brought upon by winter under ice. With their wide ranging diet, Daphnia are a dominant competitor of several species of zooplankton, including the rotifers (Gilbert and Stemberger 1985; Gilbert 1988a). The impact of Daphnia on the rotifer populations is two-fold. These cladocerans can affect rotifers through resource competition for the small size particles (Gilbert 1985; Gilbert 1988a; Jack and Gilbert 1993). Large Daphnia have a major size advantage over rotifers (see 15 Figure 1.3 Large prey: Daphnia rosea (Cladocera: Daphnidae). Suspension feeders on 0.5 - 60 um particle range Daphnia rosea Female with young in brood chamber Body size: 1200-1600um, up to 2000 u.m (excluding spine) Figure 1.4) and can also affect them directly by preying on them, or by battering them within their "filtering" appendages while feeding on other food particles (Gilbert and Stemberger 1985; Burns and Gilbert 1986a; Gilbert 1988a). In the latter case, the rotifers are not necessarily ingested. They can be rejected with the feeding water current or in the bolus with other rejected particles. Although the rotifers might not be killed immediately by the process, they can be gravely injured (Gilbert and Stemberger 1985), which makes them more susceptible to predation, reduces their ability to feed, or prevents their reproduction. All of this leads to reduced rotifer population growth and density. The third player in this system is the phantom midge Chaoborus (Diptera: Chaoboridae) (see Figure 1.5), a predatory aquatic larva which changes body size and mouth gape size between each of four larval instars (see Figure 1.6). Young instars are limited to small prey sizes, while older instars can feed on increasingly larger prey. Young instar larvae can grow fast, but are prone to starvation when prey of the right size category are rare. They have a relatively short life span, for example, one to two weeks in the first instar (Neill 1988a). Older instar larvae grow slower but can survive without a meal for a longer period of time than the younger instar larvae. These differences in metabolic characteristics and in diet of each instar can lead to different community dynamics, depending on the prey assemblage, prey density available to the predator, and environmental conditions. 17 Figure 1.4 Relative size: Daphnia(on the right) versus solitary rotifer Keratella (on the lower left). 18 Figure 1.5 The predator Chaoborus sp. (Diptera: Chaoboridae): head and jaw close up. Chaoborus trivittatus fourth instar larva Variation in head capsule length between individuals : 1.5 - 2.8 mm (length from the articulation of the second antennae with the head (a) to the furthest point on head exoskeleton (b)) (a) jaw secondary antennae with chetinous teeth (used for prey capture) 19 Figure 1.6 Family portrait: Chaoborus larval instars and pupa, third instar second instar first instar fourth instar pupa Chaoborus sp. First instar: head length* 400 um gape width unavailable Chaoborus americanus Instar head* (|im) gape width + (urn 2 550 300 3 1000 430 4 1600 710 Chaoborus trivittatus Instar head* (u.m) gape width + (urn 2 700 360 3 1350 510 4 2200 850 * Fedorenko and Swift, 1972 + Fedorenko, 1975 20 Both the rotifers and Daphnia are prey for the voracious phantom midge, Chaoborus (Fedorenko 1975b; Kajak and Rybak 1979; Smyly 1980; Vinyard and Menger 1980; Pastorok 1980a; Pastorok 1981; Neill 1981b; Moore and Gilbert 1987; Riessen et al. 1988; Walton 1988; 1988a; Moore 1988b; Christoffersen 1990; Havens 1990; Moore et al. 1994). Figure 1.6 represents the different instars of Chaoborus, and the head and gape size (based on Fedorenko and Swift, 1972) for the two species present in the lake. In lakes with higher productivity, Chaoborus can produce several generations per growing season. In the oligotrophic lakes of the UBC Research Forest, this predator reproduces once a year (e.g. C. americanus) or even once every two years (e.g. C. trivittatus). Generally, emergence of flying adults occurs in late spring (May-June: Fedorenko and Swift, 1972; and this study), from aquatic pupae floating at the water surface. They mate and females lay egg rafts on the water surface. The eggs hatch within a few to 48 hours, depending on temperature (pers. observ.), into small first instar larvae, which feed mainly on rotifers and other small-sized prey (Moore and Gilbert 1987; Walton 1988; Moore 1988b; Havens 1990; Moore et al. 1994) because of their small gape size. Within a few days, the first instars molt to second instars which also feed on small prey, including small copepod nauplii (Moore and Gilbert 1987; Moore 1988b; Moore et al. 1994). When the second instars molt to the third instars, they can feed on larger prey including young Daphnia, other small cladocerans, copepodites and adults of small copepod species. By late summer, most larvae reach the last, largest and most voracious fourth instar. At this stage, the diet spans a wide range of prey species and sizes, from small rotifers to larger 21 prey such as Daphnia (1 mm in size head to base of tail spine) (Fedorenko and Swift 1972), other cladocerans and copepods (Riessen et al. 1988). Chaoborus overwinter mainly as fourth instar larvae. In the spring, they feed before pupating, and metamorphosing into flying adults to start the new generation. For species that take two years to complete their life cycle, such as C. trivittatus, the larvae spend a second summer in the lake as fourth instar larvae, and only metamorphose during the second spring. With their longer life history, the predators represent a slow variable while the fast growing Daphnia and rotifer populations represent the fast variables in the dynamics of this zooplankton community. The thoughtful synthesis of the direct and indirect interaction between Daphnia, rotifers and Chaoborus was started by Neill and Peacock (1980) and the community dynamics were elaborated by Neill (1981b; 1984; 1985; 1988a; 1988b). I represent these dynamics in schematics (Figure 1.7 and Figure 1.8). By late spring, when Daphnia blooms, the rotifer populations decline. This reduces the density of small prey available to the newly hatched young instars of Chaoborus (Figure 1.7). Drastic Chaoborus larval mortality occurs during the first and second instar stages (Neill 1988a). Because the number of third and fourth instars is low, Chaoborus have a limited impact on the population of Daphnia, which then dominate the dynamics of the whole community, including that of the predator (Figure 1.7). The main interaction in this system is one of competition between the suspension feeders. Predation plays only a minor role. This configuration of 22 Figure 1.7 Schematic of interactions in the Daphnia domain of attraction: the competition state zooplankton community interactions Size of arrow represents relativeimpact of A on B (see text for explanation) ROTIFERS competition predation recruitment CHAOBORUS instars 1 & 2 DAPHNIA i predation CHAOBORUS instars 3 & 4 Main interaction: COMPETITION Daphnia depresses rotifer population Stability: persistence and resilience 23 Figure 1.8 Schematic of interactions in the hypothesized Chaoborus domain of attraction: the predation state zooplankton community interactions Size of arrow represents relativeimpact of A on B (see text for explanation) ROTIFERS . . DAPHNIA competition recruitment CHAOBORUS instars 1&2 CHAOBORUS instars3 &4 Main interaction: PREDATION Chaoborus depresses Daphnia and rotifer populations Stability: persistence? (need to be tested) 24 the food web is persistent over several generations, and is resilient to even large disturbance such as substantial Daphnia biomass removal (Neill 1985). This is what I refer to as the Daphnia domain of attraction, the "competition state" of the community. Under experimental conditions, a different state has been observed (Neill 1988a; 1988b). Through reduced competition between the suspension feeders caused by Daphnia reduction, or through increased predation on Daphnia, the rotifer populations bloom over a longer period of time. If this takes place during the time when Chaoborus is reproducing, the recruitment of the young Chaoborus instars is improved (Figure 1.8). More of them grow into third and fourth instars by the end of the summer. When the fourth instars are numerous, their impact on prey populations is large. The predators can now reduce the population of Daphnia, and this results in reduced competition on the rotifer populations which remain high (Neill 1985; 1988a; 1988b). Thus, by the fall, the community is switched from the Daphnia domain to a new state I call the Chaoborus state, where the main interaction in the system is now predation instead of competition between the prey. Here I use the expression "Chaoborus state" rather than "Chaoborus domain" because the persistence over several generations of this new state still remains to be tested. The work described above revealed evidence satisfying the first criterion, the presence of a threshold, and partially satisfying the second criteria, persistence of 25 the "competition" state. No evidence exists that the alternative "predation" state can persist in oligotrophic conditions. If this new state could be shown to persist over several predator generations, the community would have two domains of attraction. On the other hand, if the predation state does not persist it would instead represent a transient dynamic. Such transients, especially when relatively long, can play an important role in ecosystem dynamics. Here, I test the hypothesis that the zooplankton community of an oligotrophic lake has two domains of attraction, the Daphnia domain and the Chaoborus domain. In particular, I carry out an inter-seasonal experiment in large enclosures to test whether the zooplankton community can persist in the "predation state" first, over the winter, and second, over several predator generations. Persistence of the "predation state" requires that four assumptions, which form the key research issues in this thesis, are met (Table 1.1). The first two concern predator dynamics. First, recruitment of young predators in the summer should be enhanced in the Chaoborus state compared to their recruitment levels in the Daphnia domain. Second, fourth instar predator larvae should be able to survive overwinter at high density so as to continue to suppress the Daphnia population when these cladocerans hatch from resting eggs in the spring. Ability of the predator to survive overwinter was tested in laboratory experiments as well as in 26 Table 1.1 Requirements for the persistence of Chaoborus state as an alternative domain of attraction to the Daphnia domain Predator dynamics • Young predator instars must recruit in higher density in summer in Chaoborus state than in Daphnia domain • Fourth instar larvae must be able to survive at high density over the winter Prey dynamics • Lower Daphnia density prior to young predator recruitment • Higher solitary rotifer densities prior to young predator recruitment 27 the field. I will test these two criteria for predator dynamics in Chapter 2. The next two elements involve prey dynamics. Third, Daphnia population density prior to the young predator recruitment period should be lower in the Chaoborus state than in the Daphnia domain. Lastly, in response to the low Daphnia density, solitary rotifers should be released from competition and be found at higher densities prior to the young predator recruitment period in the Chaoborus state. I will test the response of the prey in relation to each other and in relation to predator densities throughout the predator life cycle in Chapter 3. Analysis of field data for systems with thresholds in their dynamics might require new or combined methods. Already in 1966, Morley (1966a and b, cited in Noy-Meir 1975) had pointed out that conventional statistical analysis of "average effects" would mask the discontinuity in the results, and the theoretical insight derived from them. Scheffer (1998, p. xvii) went further and stated: "The classic ideas about hypothesis testing are of rather limited use in ecology." He explains that this is because strong inference assumes that competing hypotheses are general and mutually exclusive. However, in ecosystems, several independent mechanisms can contribute to an observed phenomenon. Moreover, one mechanism can dominate, but its dominance will differ from case to case and may even shift in time. In Chapter 2 and 3, I use linear regressions and correlations to underline the problems generated in data interpretation when these techniques are used for data sets including alternative states. 28 During my field work, I found that water temperature in Shirley Lake could have a greater influence than first anticipated on the predator and prey, and on their interactions. Information on the role of cold temperature in the spring on Daphnia population dynamics is included in Chapter 3. Laboratory experiments on the effect of cold temperature on the development and survival of the predator revealed important differences between the two species of Chaoborus present in the lake. These differences could affect the persistence of the predation state. The results of those experiments are presented in Chapter 4. Finally, in Chapter 5, I link the information gathered from different scales of observation from Chapters 2, 3 and 4 to discuss the mutual impact of the predators and the prey on each other. I discuss also the impact of different disturbances on the potential for multiple domains of attraction in Shirley lake zooplankton community, and in other lakes at similar and different nutrient levels. Other biological systems show the presence of thresholds in their dynamics. Determining if these thresholds lead to alternative domains or to transient states could be important for ecological management. I conclude that using a variability perspective rather than a focus on equilibrium, allows us to include in our repertoire of explained community behaviors the discontinuous observations which have so far been discarded as exceptional events. 29 CHAPTER 2 TRANSIENT STATE OR DOMAIN OF ATTRACTION: TESTING THE PERSISTENCE OF THE CHAOBORUS STATE 2.1 INTRODUCTION Fishless lakes containing the predator Chaoborus, the phantom midge, show great variability regarding the impact of this aquatic insect larva on the zooplankton community. In eutrophic lakes, the phantom midge Chaoborus can have a large impact on the zooplankton community structure and dynamics (Kajak and Ranke-Rybicka 1970; Hillbricht-Ilkowska et al. 1975). In lakes where Chaoborus are numerous, the community is dominated by smaller zooplankton species, such as rotifers and Bosmina, while Daphnia are low in density or absent (Havens 1990). In oligotrophic lakes on the other hand, Chaoborus was found to have little impact on the zooplankton community (Neill 1981b; 1988a) unless Daphnia was absent (Stenson 1990). When present in large numbers, Daphnia dominate the dynamics of the community (Figure 1.7). My goal is to determine if these communities, the one where Daphnia dominate and the one where Chaoborus does, are two different and independent entities with separate structure and dynamics or if they are two facets of a single community which can switch between alternative domains of attraction. Enclosure experiments using a gradient of disturbances have shown that in environments with higher nutrients than found in oligotrophic lakes, 30 Chaoborus can have a large impact on the zooplankton community (Neill and Peacock 1980; Neill 1988a). These graded experiments suggested that the change from Daphnia to Chaoborus state is discontinuous, that is the response of the system is disproportionate with the change applied to the system. The zooplankton community can be, at least seasonally, switched to a state where predation rather than prey competition dominate the community dynamics. Persistence of the predator state requires that the starvation-prone young predator instars overcome a recruitment bottleneck generated by size-structured prey competition. Even when young predator recruitment is improved, the predator state can persist only if the older larvae, now in higher densities than in the Daphnia domain, can survive through the low food period created by winter conditions. The predators must survive in sufficient density to prevent rapid Daphnia population growth in the spring and to allow rotifer populations to multiply prior to the next predator recruitment period. Thus, a feedback loop can be established which could promote persistence of the new state as an alternative domain of attraction to the Daphnia domain. The mechanisms leading to the change in community state and which led to the hypothesized conditions required for persistence were uncovered through within-season (i.e., within predator cohort) experiments (Neill and Peacock 1980; Neill 1981b; Neill 1984; Neill 1985). However, only experiments monitoring changes over several predator generations (i.e., between cohorts) can address the 31 persistence of the predator state. To test the hypothesis that the Chaoborus state is an alternative domain of attraction for the zooplankton community, I used field enclosure experiments to create a gradient along a disturbance strength axis. These experiments allowed me to assess the presence and relative location of the threshold in community dynamics and to test the persistence and resilience of the Chaoborus state. Daphnia density reduction and predator additions generated a pulsed disturbance on the community. I monitored community responses in the enclosures and the lake over three years using a variable sampling interval to address community dynamics at different time scales through the seasons. As a complement, I used a laboratory experiment to better evaluate the ability of the predator to survive in cold temperature at low food level (winter conditions). In this chapter I address the impact of the enclosure experimental treatments on the predator dynamics. 2.2 MATERIAL AND METHODS I briefly present the main methods used in this study. The detailed descriptions are presented in the sections below. Experimental enclosures were set in Shirley Lake, B.C. (see section 2.2.1) to test the presence of the alternative predator state and to determine its persistence. Experimental treatments included partial Daphnia removal, predator additions and nutrient additions (see section 2.2.2). The lake and the enclosures were monitored for three years during the ice-free season following an adaptive sampling design where the sampling pace varied 32 depending on predator developmental stage and water temperature (see Section 2.2.3). The zooplankton community was sampled by hauling vertically a 102 ii m mesh zooplankton net for the predator and large zooplankton such as Daphnia, and a 50 um mesh zooplankton net for rotifers (see section 2.2.4). Phytoplankton was sampled using a small diaphragm pump to collect water from 1 m layers from the surface down to 3 m (see section 2.2.4). Temperature was sampled on each sampling date while water for nutrient analysis was collected twice per year in June and in October (see section 2.2.4). Techniques for counting predators are detailed in section 2.2.4 while those for counting zooplankton and rotifers are presented in section 3.2.2. Predictions based on the experimental design are presented in section 2.2.7. 2.2.1 Field site and enclosure design Shirley Lake (Figure 2.1) is a small, fishless, oligotrophic, mountain lake (N 49° 20' 42", W 122° 33' 37") in the Malcolm Knapp Research Forest (Maple Ridge, B.C.) of the University of British Columbia (Vancouver, B.C.). The composition of its zooplankton community is similar to that of its close neighbor Gwendoline Lake (where the community interactions had previously been worked out) prior to 1982, after which fish introduced in 1979 had modified the zooplankton community. The zooplankton community in the limnetic zone of Shirley Lake is composed of several species of rotifers, cladocerans, copepods, and aquatic insects. The rotifers included several genera of solitary rotifers, Keratella 33 Figure 2.1 Map of Shirley Lake in the Malcolm Knapp UBC Reseasch Forest, Maple Ridge, B.C., Canada (after Butler, 1990) and Kellicottia (Brachionidae), and Polyarthra (Synchaetidae) (Figure 1.2), and one colonial genus, Conochilus (Chonochiloidae). Cladocerans included several different genera of which Daphnia (Daphnidae) (Figure 1.3) was found to be the predominant one. Holopedium appeared regularly, sometimes in high numbers, but seemed to play a minor role in the community functioning as a weak suspension feeder and competitor. It was well protected in its gelatinous coating. Bosmina (Bosminidae) and Diaphanosoma (Sididae) were found to be rare in the lake. Copepods included mainly two species of calanoid copepods, Diaptomus kenai and D. leptopus (Diaptomidae), with the occasional presence, at low densities, of a cyclopoid copepod, Cyclops (Cyclopidae). The main aquatic insect larva found in the samples was the dipteran Chaoborus (Chaoboridae), also called the phantom midge. Notonectids (Hemiptera: Notonectidae), and the occasional diving beetle (Coleoptera: Dysticidae) were also found in some samples. No fish occurred in Shirley Lake and salamander larvae (Ambystoma gracile) were present. The generalities about the life style and life history of the rotifers, Daphnia and Chaoborus, the main players in the community dynamics, were described in Chapter 1. Details of the life histories of the predators specific to Shirley Lake are described below. In Shirley Lake, the life history of C. americanus (Figure 2.2 A) followed a pattern with pupation in June (Figure 2.2 B) and one generation per year, similar to that described above for studies done previously on neighboring lakes in the Research Forest (Teraguchi and Northcote 1966; Fedorenko and Swift 1972; Neill 35 Figure 2.2 Life histories of Chaoborus americanus and C. trivittatus in Shirley Lake from spring 1992 to spring 1995 Fourth instar Third instar Second First instar instar (both species) Pupation timing ^- americanus C. trivittatus 1992 MJ JASON 1993 AM JJASON 1994 1995 AMJJ A SON .... AM J i t <i C. americanus •i >i 'i • i. « i »i | :i ( A 'I\ ..JI. 120 210 300 390 480 570 660 750 840 930 1020 1110 1200 1290 _j i * • • i— 1st instar MJJASON AMJJASON AMJJASON AMJ 120 210 300 390 480 570 660 750 840 930 1020 1110 1200 1290 1992 1993 1994 1995 36 1988a; 1988b). C. trivittatus, however, showed a different life history than observed in those lakes. In Shirley Lake, C. trivittatus had two pupation periods per year, in early spring and in mid-summer (Figure 2.2 B). This generated an unusual life history with more than one peak per instar, and with second and third instars present overwinter in addition to the fourth instars (Figure 2.2 C). I have not been able to determine if C. trivittatus in Shirley Lake has a one-year or two-year, or an even more complex life history. The life history of C. trivittatus, although unusual, leads to the presence of fourth instar larvae year round just as in the lakes with the two year life cycle. In Shirley lake, I set up field experiments in large enclosures where I generated a series of different initial conditions. I gathered information on inter-year dynamics, such as predator overwinter survival, and on intra-year dynamics such as early summer young predator recruitment. I installed ten enclosures (Figure 2.3) in the spring of 1992 before the start of the reproductive period for the predatory aquatic insect, Chaoborus. The enclosures were translucent and made of 4 mil woven, white polyethylene plastic. They were sewn by False Creek Industries (Vancouver, B.C.), into cylindrical bags with a bottom sheet to close them off to the sediment (see Appendix A for details of enclosure design). The top of the enclosure (first 1 m) was made of yellow-coated polyethylene plastic to better resist UV light damage which, otherwise, renders the plastic brittle and prone to breaks. The yellow plastic was sown into a sleeve 37 Figure 2.3 Enclosure set up and experimental design Experimental set up diagram High predation Low nutrient Low predation High nutrient Low predation Low nutrient •..Lake station t' High predation Low nutrient High predation High nutrient 38 •.Lake station 2 Medium predation Low nutrient p Low predation High nutrient Low predation Low nutrient Medium predation Low nutrient Lake station 3. High predation High nutrient in which an extruded polyurethane plastic float (17.5 m long X 15 cm high X 5 cm wide) was inserted to keep the enclosure afloat in the lake. Each enclosure (5.6 m in diameter, 4 m deep) contained approximately 95,000 litres of water. Each enclosure was surrounded and attached to an hexagonal 2X4 inches cedar wood frame to protect them from floating debris, and allow boat docking without damaging the enclosures. The enclosures were installed in two side-by-side columns in the middle of the lake (Figure 2.3), maintained in place by ropes tied to shore and by cement block anchors on the bottom of the lake. Unfiltered water from the lake was pumped into the enclosures by five diaphragm pumps with 7.5-cm diameter hoses. Each pump was used to fill two adjacent enclosures in alternation over several hours to provide equivalent mixtures of zooplankton (see Appendix B for details). After filling up, enclosure communities were allowed three weeks to acclimate, before the predator additions were started. 2.2.2 Experimental design Experimental treatments included first Daphnia reduction, followed by predator enhancement and nutrient increase to insure that a Chaoborus state could be generated in at least some of the enclosures. The goal was to investigate the conditions required for persistence of the Chaoborus state over one or more continuous Chaoborus generations. Simply by pumping water containing the zooplankton community into the enclosures, I achieved in all of them a 39 fortuitous but substantial Daphnia reduction (Figure 2.4). These reduced Daphnia populations, lower than in the lake, were allowed to grow for three weeks and were subsequently suppressed further by additions of different levels of predators and nutrients to create the final experimental design. Within the predation and nutrient addition experiments (see Table 2.1), I used different levels of treatment to explore aspects of the resilience and resistance of the system and to try to locate the boundary between the Daphnia (competition) domain and the Chaoborus (predation) state. I randomly assigned each enclosure to a treatment combination (predation X nutrient) using one of three predation levels (low, medium, high), and two nutrient levels (low, high) (Table 2.1). High and low predation treatments at both levels of nutrients were replicated. If replicate treatments (predation X nutrient) were selected for the same column in the side-by-side enclosure set up (Figure 2.3), the draw was redone to assign a new location. Daphnia reduction and predator additions were one time pulse disturbances to create different initial conditions. Nutrient treatments were repeated additions to maintain two different productivity levels. The enclosures were then monitored for three years without restarting the experiments. Predation levels were set once at the start of the three-year experiment by adding different numbers of fourth instar Chaoborus larvae. To set up the required predation levels, I collected predators at night from the lake by towing a Wisconsin type net with aim diameter mouth opening behind a boat with a 3-horse power motor. The animals were stored in 20 litre carboys. They were sorted 40 Figure 2.4 Daphnia reduction: densities of Daphnia in enclosures after water pumping and before predator additions. Density in lake for the same time period is provided for reference. 103 Daphnia CO I g •i ioM T3 -a 101 J 10°. 9 .ft: ft^ ft' fti fti fti ft' ft' Lake B D ft" si ft"?l ft' ft' fti ft' fti fti G H I | J pumping pairs 41 Table 2.1 Experimental design for predator and nutrient additions. Predation ( X lake level): 2-3 6-8 13-16 LOW MEDIUM HIGH Fertilization: C, J A, D (0.02ug P / L; N:P =30)* LOW F (0.2ug P/ L; N:P =30) HIGH B, G I E, H * added at start for a few weeks; nutrient additions were then stopped to maintain nutrient level close to lake levels 42 in the lab to remove most prey, and kept cold till they were returned to the field site, within 60 hours from capture. The water containing highly concentrated Chaoborus densities was split at lake side into equal amounts and added to the different enclosures until the different predation levels were obtained. The entire process, from Chaoborus captures and sorting to introductions, was accomplished over a 9 day period (May 14 - 22,1992). Once fourth instar predators had been introduced to the enclosures they were allowed to finish their life cycle without interference. Note that recruitment of first instars early each summer was done naturally. Aerial adults from the lake and surrounding area, not only those emerging from within an enclosure, contributed the eggs for the new cohort in this enclosure. Egg deposition was assumed to be random between enclosures. I personally observed numerous egg rafts deposited on the water surface in all enclosures and on the lake. There seemed to be no preference on the part of laying females for enclosures with good rearing conditions. For example, eggs were laid even in enclosures where, over time, prey populations failed altogether. Thus, larval recruitment of instars other than the first instar in an enclosure is most likely due to higher larval survival and not higher egg deposition from emerged adults of the previous enclosure generation. The fertilizer additions were started in July 1992 and were repeated regularly. The aim was to maintain two different levels of nutrients throughout the 43 experimental period: low nutrient levels as in Shirley Lake and high nutrient level, above the Shirley Lake nutrient concentration. The low nutrient enclosures received additions of potassium phosphate (KH2P04) and sodium nitrate (NaN03) based on adding 0.02 ug P l"1 week"1 in an atomic ratio of 30 N:P. The low nutrient enclosures received weekly nutrient additions till the end of July and then they were not fertilized again in an attempt to keep the nutrient levels as low as in the lake. The high nutrient treatment received weekly or bi weekly fertilizer additions (depending on the sampling interval) based on 0.2 fig P l"1 week"1 in an atomic ratio of 30 N:P. Both fertilization levels are relatively low on the oligotrophic to eutrophic scale. Based on a pilot experiment I did in the year previous to this experiment, I found that the use of nutrients at a concentration of 2 (ig P l"1 week"1 (atomic ratio of 30 N:P) would generated blue-green algae bloom and the collapse of the zooplankton community. To avoid this undesirable state, I limited the nutrient levels to the levels mentioned above. In addition to the enclosures, three permanent lake stations were sampled and their location was marked with a float anchored to the bottom of the lake (Figure 2.3). I will report here only the results for the deepest station, Lake Station 2. Because Shirley Lake is small, the three station were close to each other (25-30 m apart). No difference in species composition was recorded in samples where all three station were counted. Qualitative differences observed in zooplankton density were temporary and generally occurred at the onset of the increase phase 44 of a species population dynamic in the spring. Station 2 was sampled from 0-3 m to compare with enclosure samples and from 0-6 m to record additional information on Chaoborus life history in Shirley Lake. 2.2.3 Variable sampling interval: towards an adaptive sampling design Chaoborus, zooplankton, rotifers, phytoplankton and water temperature were sampled between May 1992 and June 1995. Nutrient samples were taken in June and in October each year. The ice-free sampling season started in March/April and ended in October/November (see section 2.2.4 for Chaoborus field sampling methods and section 2.1.1 for other field sampling techniques). I chose a sampling design with a variable sample interval for three reasons. First, slow dynamics (long time scales) can be followed using long sampling intervals while fast dynamics (short time scales) require short sampling intervals. Zooplankton communities have the potential to respond nonlinearly to even small disturbances in environmental or biotic conditions, thus dynamics could suddenly change gear from slow to fast. Second, generation times between species can differ resulting in dynamics functioning at different time scales, often simultaneously. Third, under the influence of environmental factors (e.g., temperature), generation time within a species can vary from season to season or lead to population dynamics (e.g., Daphnia population bloom, or Chaoborus reproduction) taking place on different dates from year to year, both of which require distinct sampling intervals rather than an average regular sampling 45 interval. I adapted, for my study, a general design with variable sampling interval as suggested by Ouimet and Legendre (1988). The sampling interval varies depending on the levels observed for a specific variable such as presence or absence of certain species or instar, or such as chosen levels of biomass for a species or of an abiotic factor. Thompson (1992) refers to sampling designs with variable pace as "adaptive sampling". He emphasized that this type of sampling design maximizes the information gathered and minimizes the costs, while allowing one to follow organisms or dynamics which are variable in time and/or space. Adaptive sampling designs allow efficient sampling of systems with a potential for nonlinear responses, however the rules pointing out which sampling interval to use should be decided a priori. I designed my sampling plan as follows. First I chose a minimum systematic sampling interval: once per four weeks, which I refer to hereafter as the "monthly" sampling. Second, I varied the sampling interval I used depending on the period in the life history of Chaoborus, and on temperature. The monthly samples included daytime sampling of all the variables (water temperature, phytoplankton, rotifer, zooplankton and Chaoborus), and night sampling for Chaoborus older instars. Third and fourth instar Chaoborus vertically migrate (Teraguchi and Northcote 1966; LaRow 1968; Fedorenko and Swift 1972; Swift 1976). Descending deeper as light increases, they spend the day near the bottom of 46 the lake, while at night they gather and feed in the surface layers, at which time I sampled their densities more accurately. The monthly night samples were collected on the same date as the daytime samples. The initial monthly sampling date for each year was chosen in such a way that most monthly sampling dates for that year would fall closer to the new moon than to the full moon. Sampling intervals were shortened during Chaoborus reproduction period and young instar growth, and as temperature warmed up in the spring and early summer. I lengthened the interval in late summer and fall as water temperature cooled down and community dynamics were expected to slow down. During Chaoborus emergence, and during the period with first and second instar stages, dynamics take place at a faster pace and sampling was adjusted accordingly. These high intensity samples were collected only during daytime because the young Chaoborus instars do not migrate at that time of the year(Goldspink, 1971; Parma, 1971; Fedorenko, 1975a; Pastorok, 1981; but see comments in Appendix C on young instar reverse diel migration in the fall). I sampled the daytime categories mentioned above weekly during Chaoborus pupation, semi-weekly (twice per week) during the first instar period, and back to weekly during the second instar phase. During the third and fourth instar stages, I sampled bi-weekly or monthly depending on water temperature. In 1992, the first year of the experiment, I sampled weekly through the entire season, and semi-weekly during the Chaoborus reproduction period, to record a more 47 detailed picture of the seasonal dynamics and acquire basic information on the behavior of the community in Shirley Lake. The adaptive sampling design ensures that samples are collected in relation to dynamic events rather than to the calendar, and that important events are not missed, even if for example they are delayed by environmental factors. This means that not all of the samples collected need to be counted. Again I used an "adaptive" strategy to choose the samples to be counted (see section 2.2.4 for details of the counting method). One replicate sample for the full series of monthly samples were counted. Furthermore, for June and October of each year, the second replicate monthly samples were counted. Where change in density was large, bi-weekly, weekly and/or semi-weekly samples were counted as required. Due to time limitation, I was not able to implement a full adaptive sampling strategy for counted samples. 2.2.4 Sampling methods, identification, and counts for Chaoborus: Below I describe methods relating to Chaoborus sampling, identification, and counting. I detail the methods relating to zooplankton, and rotifer sampling and counts in Chapter 3. Chaoborus were collected, in the lake and in the enclosures, by hauling a Wisconsin type zooplankton net (mesh size: 102 \im; mouth diameter: 0.4 m) vertically from 3 m to the surface. Mean sieving efficiency for the net was 48 determined in 1992 with an electronic flowmeter outfitted on the net rim with which ten readings were made and averaged. The zooplankton net was new and sieved at a mean efficiency of 95%. The volume sampled was approximately 377 1. Complementary samples were taken on some dates where I collected hauls, at the deep lake station 2, from 6 m to the surface. Day samples were generally taken between 10:00 and 13:00, while night samples were generally collected between 23:00 and 1:00 Pacific Daylight Saving time. Samples were preserved in 100 ml glass jars with plastic lids, using 5% sugared formaldehyde solution. Chaoborus identification was done under a Wild M5 stereoscope at 12X to 40X power depending on the instar, and under a Nikon inverted microscope at 100X for pupal species identification using Saether (1970) and Borkent (1979). To identify larvae to species, I used Fedorenko and Swift (1972). First instar larvae are indistinguishable between species (Fedorenko and Swift 1972). Counts were done in grided petri dishes (Edmonson 1971b) under a Wild M5 stereoscope. Chaoborus third and fourth instar larvae were counted at 12X in entire samples. Samples with large numbers of first and second instars were split using a 250 ml Folsom plankton splitter (splitting wheel) with 4 divisions (4 X 1/4 subsamples) (Edmonson 1971b). They were then counted at 12X and/or 25X. The data were used to produce a three-year time series for each larval instar of each Chaoborus species, data on which results from section 2.3 are based. I 49 present the detail of these time series and some idiosyncrasies of Chaoborus dynamics through time in Appendix C. 2.2.5 Laboratory experiments I set up laboratory experiments to investigate the overwinter survival abilities of fourth instars of Chaoborus americanus and C. trivittatus with and without food. Hypothetically, if the predator can survive overwinter at high density, the top-down signal can be passed from one predator generation to the next, and the predation state can persist as a potential domain of attraction. Just before the lake froze at the end of November 1992, I collected Chaoborus fourth instar larvae from Shirley Lake, at night, using the 0.4 m diameter Wisconsin type zooplankton net with a 102am mesh size. I stored the collected larvae in environmental chambers at 5°C, in 20 1 plastic containers filled with lake water containing a high density of zooplankton. After two days, I selected individuals with visible food in their crop or gut. I set the larvae individually in 150 ml plastic containers filled with lake water filtered through 20 um mesh netting and assigned them randomly to "FED" or "NOT FED" treatments. I put all the larvae in a single environmental chamber at 5°C, in the dark (0 hr. light : 24 hr. dark) to simulate winter conditions under ice in the lake. I recorded the status of each larva weekly: alive or dead, with/without food in crop or gut, and its life stage (larva, pupa, adult). In the "Fed" treatment, I also recorded the number of dead and/or eaten prey, and replaced them with fresh prey. Prey 50 consisted of a variety of nauplii, small Daphnia, other small cladocerans, copepodites or adults of Diaptomus leptopus, depending on availability. I terminated the experiment when the last larva finally pupated, 70 weeks from the start of the experiment. 2.2.6 Initial experimental conditions and general seasonal patterns in enclosures Predator additions, which included fourth instars of both C. americanus and C. trivittatus, were finalized on May 22, 1992. The first night samples, to evaluate initial predator experimental levels and monitor the system's responses, were taken on May 26, 1992, four days later (Figure 2.5). I averaged the density measurements of two night sample replicates for each enclosure and for the lake station. Fourth instar Chaoborus density in Shirley Lake was 42,individuals m"3. Mean densities for the four low predation enclosures varied from a minimum of 84 to a maximum of 139 fourth instars m"3 (2-3 times lake density). Mean densities for the two medium predation enclosures were 289 and 361 fourth instars m" (6-8 times lake density). The four high predation enclosures showed, as expected, the highest mean densities and the greatest variability, with levels varying from 549 to 703 fourth instars m"3 (13-16 times lake density). Predator density between predator treatment levels were distinct (ANOVA on the logarithm of the density: d.f. (2,7); p«0.05). A clear predator density gradient was thus set (Figure 2.5). 51 Figure 2.5 Initial densities of fourth instar predators in Shirley-Lake and in the enclosures on May 26th, 1992. Data points represent means of two replicate night samples per station. CO c X! o CO co "° e: g S 3 .S -t-» —^ CO o fin 1992 Predator additions ^ Low Medium High 103 -i 10 • • • ~i I—I—I—I I—I 1—i—i—r L2CJ BGF I ADEH Lake station and enclosures 52 The combination of Daphnia density reduction through pumping and the predator density gradient through additions of fourth instars resulted in Daphnia densities lower by 65% to 99% in enclosures than in the lake at the beginning of the experiment (Figure 2.6). C. americanus life history pattern in the enclosures was roughly the same as in the lake (see Section 2.2) with a single reproductive period per year in the summer. However, C. trivittatus did not recruit well in the enclosures. This species did not survive past mid-summer 1992, leaving C. americanus as the main predator in the enclosures. Thus, when comparing total numbers of predators, this total includes both C. americanus and C. trivittatus for the lake data, while in the enclosures, the total number of predators includes both species up to mid-1992 and thereafter includes only C. americanus. Over the three year monitoring period, 1992 to 1994, different enclosures maintained their community for different duration (Figure 2.7). Two enclosures, A and F, which had been filled with the same pump (see Appendix B), failed within two months (mid-summer 1992) of the start of the experiment. An enclosure was deemed to have failed when it lost its predator cohort, i.e. when the density of predators in the enclosure was much lower than the lake predator density. Enclosures A and F did not recruit a new predator cohort because of prey population failure. I do not use the results from these two enclosures in analysis. 53 Figure 2.6 Large prey density reduction: percentage of Daphnia density removed from enclosures relative to Daphnia density in the lake after predator additions (May 26th, 1992) Fvl Low predator additions H Medium predator additions | High predator additions CJBGFIADEH Enclosure 54 Figure 2.7 Duration of Chaoborus populations in experimental enclosures and in Shirley lake. CO c o CO CC "D a> a. X a X Q UJ High Nut. Low Nut. § HighNut. I Low Nut. F High Nut Low Nut. Lake i i i i i MJ J A SON 92 -mr A MJ J A SON 93 94 55 Three of the four low predation enclosures (C, B, G), the medium predation enclosure (I) and one high predation enclosure (D) had functional Chaoborus populations until mid-summer 1993 (Figure 2.7). In enclosure J (low predation), the Chaoborus population dwindled in early summer 1994. Two high predation Chaoborus populations (E, H) lasted till early and mid-summer 1994 respectively. Chaoborus lake populations persisted through the whole study period. Note that enclosure failure most often occurred in mid-summer, during the young instar recruitment (Table 2.2). This was usually accompanied or preceded by failure in prey populations and by visual cues such as proliferation of blue-green and colonial green algae in sufficient densities as to reduce light penetration depth and reduce the sieving capacity of the sampling nets. Thus, in the results section presented below I make use of different enclosures to address the community dynamics over different periods of time. 2.2.7 Predictions based on experimental design I test the hypothesis that the zooplankton community in Shirley Lake can function in two different domains of attraction. The zooplankton community in the enclosures should be able to switch from the Daphnia domain, as found in the lake, to the Chaoborus state (see Section 2.2, for details of community functioning). The latter must persist for several predator generations to be declared an alternative domain. If persistence is not achieved, Chaoborus state is considered a transient state of the Daphnia domain, i.e., a non-persistent state with a different functioning than that of the Daphnia domain. 56 Table 2.2 Timing of enclosure recruitment failure YEAR at Cl stage at C2 stage at C3 stage at C4 stage 92 — A, F — — 93 — CAGJ B — 94 — E,H,J — — The switch to Chaoborus state requires improved young predator recruitment in relation to decreased Daphnia densities and enhanced solitary rotifer densities prior to the young predator recruitment period in the summer (Table 2.1). The persistence of the state necessitates that fourth instar larvae possess the ability to overwinter in high densities such that they can prevent rapid population growth of Daphnia early in the season. Experimentally, I imposed a predator density gradient on the enclosure communities and monitored the transmission of this signal through the different life history stages of the predator and the related prey dynamics. I made the following predictions for the predator and prey dynamics (Table 2.3). In terms of predator dynamics, under the two domain hypothesis, I expected that more fourth instar predators in the spring would generate more young recruits in the summer and more fourth instar larvae in the subsequent spring, not because of more eggs but because of better survival. Thus, I expected for both size categories that densities would be higher in the enclosures than in the lake. I also expected higher densities of predators in the high predator treatments than in the low predator treatments in relation to the imposed predator density gradient. In other words, correlations between the predator gradient and the recruitment of young instars in the summer and the survival of fourth instars until the subsequent spring should be positive. 58 Table 2.3 Predictions for the relationships in predator densities and in prey densities between the lake and the enclosures, and between the low and high treatments in relation to the experimentally imposed predator gradient. Hypothesis: Under reduced Daphnia density and enhanced predator densities, all enclosures are expected to cross the threshold and switch to Chaoborus state Variable PREDICTIONS Lake vs. Enclosures Enclosures: Low vs. high Correlation with predator gradient Predator Dynamics Young predator recruitment density Lake < Enclosures LOW < HIGH + Fourth instar density in spring prior to recruitment of young instars Lake < Enclosures LOW < HIGH + Prey Dynamics Daphnia density prior to recruitment of young instars Lake > Enclosures LOW > HIGH -Solitary rotifer density prior to recruitment of young instars Lake < Enclosures LOW < HIGH + 59 I included here, for the sake of completeness, the predictions for the prey dynamics, although those will only be addressed in Chapter 3. I expected a negative correlation between the predator gradient and Daphnia densities, and a positive correlation between the gradient and the solitary rotifer densities. Thus, I expected to find higher Daphnia densities in the lake than in the enclosures, and in the low predator density treatment than in the high treatments. I expected the solitary rotifer density dynamics to be inversely related to the Daphnia population dynamics and to follow the same type of patterns as described for the predator dynamics above, that is higher densities in the enclosures than in the lake, and in high predator treatments than in low treatments. All of the predictions above for the predator and the prey dynamics must be met to support the hypothesis that the zooplankton community in Shirley lake has two domains of attraction based on a switch between the dominance of competition and predation processes. 60 2.3 RESULTS Predator additions resulted in substantial impacts on the community dynamics in the enclosures while nutrient additions provided divergent dynamics within treatments. For example, the time series representing fourth instar predator larvae in low predator treatments (see Appendix C-4, panel 2) are entangled throughout the season, irrespective of the nutrient treatment (Low: C, J; High: B, G). Moreover, enclosure J, a low nutrient treatment, outlasted the high nutrient enclosures (B, G), while enclosure C, its replicate lost its predator cohort early. Impacts of the nutrient additions on the enclosure community dynamics were inconclusive and were not analyzed further. Predator additions in spring 1992 affected differently the recruitment of each larval instars of the predator. I present the detailed time series based on predator densities for each instar and for each enclosure and the lake for 1992 to 1994 in Appendix C (C-l to C-4). The results presented below are based on these time series to which I refer to underline specific observations as needed. 2.3.1 Second instar predator recruitment in summer 1992: density, duration All enclosures (8/8) shifted to higher densities of second instar larvae than in the lake in summer 1992 (Figure 2.8). The enclosures recruited 4 to 11 times higher densities of second instars than the lake. Moreover, high predator treatments recruited, as predicted, higher densities of second instar larvae than the low treatment. 61 Figure 2.8 Recruitment of second instar predators in summer 1992 in relation to the density gradient in fourth instar predators in sprping 1992. Regression lines (a)with and (b) without the Lake station data point (a) log Y = 1.71 + 0.573 log X; R2 = 0.582; p=0.009 (b) log Y = 2.40 + 0.297 log X; R2= 0.429; p<0.001 Predator addition treatments Lake Low Medium High • • A • 104 Summer 1992 Second instar predator density (ind. m"3) io34 102 A • 101 102 103 Spring 1992 Fourth instar predator density gradient (ind. m ~3) 62 I present in Table 2.4 the statistical results from tests used to address the predictions presented previously in Table 2.3. I also test one more variable, the duration of the recruitment period, the importance of which I did not establish a priori. I used a binomial test to address the significance of the number of enclosures which responded differently than the lake. I used the nonparametric Mann-Whitney U test for two samples to compare densities between replicated treatment levels (low versus high). I used correlation and linear regressions to test the direction and the strength of the relationship between the predator density gradient and the chosen variable in the experiment, both including and excluding the lake data point. This allowed me to explore more systematically the presence of a difference in driving factors in the Chaoborus state. All tests were one-tailed tests with significance level set at a= 0.05. The same type of analysis was done for different stages in the life history of Chaoborus. All the tests on the second instar density data are statistically significant (Table 2.4). All predictions were met for a switch to and for the persistence of the Chaoborus state in summer 1992. Enclosures recruited higher densities of young predators and so did the high predation treatments compared to the low treatments. The linear regressions, which included low, medium and high predator treatments, with and without the lake data point, were also statistically significant, and the relation was positive as predicted. Thus, the increase in fourth instar predator larvae in the spring generated higher recruitment of 63 01 ra V) s © o « 6 Oi 45 u «-» w TJ C ra 2 £ tf) r>-( X 'C *- & 45 H_> c TJ 0) tf) cn O H u oi S tH u o> oi > tf) oi Oi S 3| W tf) w ra C > w m £ 13 tf) tf) "3 .S 01 X TJ 1° 01 45 « oi w s ^ s ° © oi -K CN OI i-H « TJ (5 ra tH tf) o CU 0) TJ (5 ra CU CU c cu cu <-H-» cu ^ co . cu >-. tH • - 5 tf) tf) £ .2 ^ c tH ^ tf) 2 £ -5 •rH CO rr< ^ TJ CU O CM CU 45 cn bo •rH TJ cu S5 be ra s5 ra 45 u cu r-o cu 45 c tf) -M cu cu •4-> cu 45 TJ CH ra cu 44 ra cu 45 c cu cu -t-> cu rQ TJ o 'tH cu OH CH CU 3 tH cu tH ra tf) CH TJ CH O CU tf) CU 45 CH o •l-H -<-> ra tH TJ CU tf) 45 p H-> tH bO tf) (5 O • rH i 1 cu i-1 +1 cu CQ tf) S5 ^ •H + ' .15 cj\ tf) ON CH ^ r£ °0 tH ^ G jS « tfi r S o bO o hJ II + +-> • rH tf) CH CU TJ tH ra H-» tf) c. TJ (5 o u cu tf) "bJD O HJ II CN ON o IT) a, o o o CO o o o o o o o o o © r>. CJ 4-> r-1 + ~. 'ATI .429 isti sul J CO 'ATI .429 j LO • s? ° rH CO o II (N II <N II N ^ CO j f>T II II j 3 o 1 rH rH <N CH CH rH 1 'tf? CH 4-> tf) tH _o o • rH 00 CU H-» 00 'tf) Ltnc tf) tf) H -rf QO tf) <u Ltnc CU 1-1 TES taile ial: ! reer •Whi ' ree ial: i cu a tH CH tH ra 6 CH o ra cu ,—1 CH CU o ra CH CO CH • rH 'H •1—1 pa CU cu cn OH DH w O o c w & In P cu cu fi p > > densit siti [IGH siti 3 O a HJ tf) 01 densit po [IGH po cu U o w o •«-> g v V e v :tion aobo AKE LOW ratio AKE -S _H LOW red red < pa young predators in the new cohort. The recruitment was proportional to the density of the fourth instar predator in the spring. Another factor indicative of an improvement in the conditions for young predator recruitment is the duration of the recruitment period. The young predators are starvation-prone (Neill 1988a). When food is limited, as in the Daphnia domain where rotifer populations are reduced by competition with Daphnia, most young predators will not live long enough to grow and switch to the next instar. The recruitment period is truncated. Thus, the period of recruitment is expected to be shorter in the Daphnia domain than in the Chaoborus state. In the lake, in 1992, the summer second instar peak for C. americanus was restricted to June (Figure 2.2, top panel). In the enclosures (Appendix C-2, panels 2, 3, 4), the peaks were wider, lasting until July, with some enclosures with low density peaks extending into August. The duration of the main recruitment period for Chaoborus americanus second instar (Figure 2.9) was longer in 7 out of 8 enclosures than in the lake (one-tailed binomial test: p < 0.035). 65 Figure 2.9 Duration of the recruitment period for second instar predators in 1992. Longer periods with high densities of second instars indicate relatively better recruitment conditions than in the lake. Long periods of low density recruitment indicate conditions where food is sufficient for maintenance but not for fast growth to the next instar. High density period —, ^ durationof second instar recruitment Low density period ^ i 1 1 1 r 0 14 28 42 56 70 84 98 112 126 140 Duration (days) 2.3.2 Fourth instar ability to resist starvation: survival in the laboratory and in field experiments 2.3.2.1 Survival in laboratory experiments at 5°C Survival of high densities of Chaoborus fourth instar larvae until the subsequent spring is a key factor in the persistence of the Chaoborus state. Fourth instar larvae must survive in sufficient densities to significantly and negatively impact Daphnia densities prior to the recruitment of the new predator cohort. In the fall, many zooplankton species, including Daphnia, retreat from the water column in the form of resting eggs. Thus over the winter, zooplankton food level availability is highly reduced for the predator Chaoborus. On the other hand, water temperature is also lower (4°C). Low temperature affected the development and survival of predator larvae (see Chapter 4). At low temperature, the predator requires lower food levels for maintenance. Low temperature in the spring also affected prey development and population dynamics (see Chapter 3). In the laboratory, I was able to test directly the ability of the fourth instar larvae to survive in cold temperature with ("FED") or without ("NOTfed") zooplankton food. The predator larvae were kept at a constant temperature of 5°C, in the dark (for methods see section 2.2.5). All but one "FED" larva survived for at least 30 weeks with a median survival time as larvae of 36.5 weeks, while "NOTfed" larvae survived a minimum of 23 weeks with a median time of 35 weeks (Table 67 2.5 A). There was no significant difference in the median time that fourth instar larvae survived with or without zooplankton food (2-tailed test: z=0.98; p = 0.327). Moreover, most fourth instar larvae were able to pupate at 5°C, with or without zooplankton prey (Table 2.5 B). Surprisingly, over half the Chaoborus trivittatus pupae were able to emerge at such cold temperature. None of C. americanus pupae emerged. The fact that both species could pupate means that development still takes place even if the temperature is very low. However, the two species must have different emergence temperature requirements (see Chapter 4). In summary, Chaoborus fourth instar larvae for both species have tremendous ability to withstand starvation in cold temperature. They can survive a median time of almost 9 months in 5°C water even when no zooplankton food is present. 2.3.2.2 Survival overwinter in field enclosure experiments In the field, 7 out of 8 enclosures maintained higher densities of fourth instar predators over the winter than in the lake (Figure 2.10). The 1:1 line on the figure indicates that densities in the lake between the two years changed little. However, in the enclosures, low predator treatments increased in density in 1993 compared to 1992, except for enclosure C which decreased to predator density similar to that of the lake. Interestingly, enclosure C had the lowest initial Daphnia reduction level (65%) and the second lowest predator addition level (89 fourth instar larvae m" ) amongst the enclosures. In medium and high 68 Table 2.5 Results from overwinter survival laboratory experiments on Chaoborus A) Number of larvae at different stages B) Time alive as larva in FED and 'NOT FED' experiments for the total number of larvae (both species combined) A) 5°C Number of larvae alive Species at start died at end pupated emerged C. trivittatus 24 4 0 20 11 C. americanus 25 2 1* 23 0 *as pupa not larva (pupated on week #69, last week of experiment) B) Alive as LARVA (larva to pupa) TIME (weeks) Min. Time (weeks) Max. Time (weeks) Total-FED 36.5 8 52 Total-NOTfed 35 23 68 Median two-sample test: Prob. > I z I = 0.3272 (S = 12; z=0.97980) 69 Figure 2.10 Chaoborus fourth instar densities in spring 1993 relation to the experimental density gradient fourth instar predators in spring 1992. Regression line for enclosures (except C) log Y = 3.41 - 0.47 log X Predator addition treatments Lake ^Low Medium High! • • A • 104 Spring 1993 Fourth instar predator density 103 (ind. m ~3) io24 io1 Spring 1992 Fouth instar predator density gradient (ind. m ~3) treatments, predator densities declined from the level that had been imposed on them experimentally, but they remained at densities 2 to 4 times higher than the lake. Table 2.6 summarizes the results of the statistical tests on the predictions made previously in Table 2.3 in relation to fourth instar predator dynamics. A binomial test confirms that enclosures maintained higher fourth instar larvae densities than the lake. A linear regression including all enclosure treatments and the lake showed a positive trend between the experimental predator gradient and the density of fourth instar predator in spring 1993 but was not significant (Table 2.6). Based on figure 2.10, enclosure C did not remain in the predator state. Enclosure C was excluded from the low versus high predator treatment comparison meant to ascertain the dynamics within the Chaoborus state. A one-tailed Mann-Whitney test. There is a significant difference between low and high predator treatments however all lowest ranks were found in the high treatments and all highest ranks were found in low treatments, contrary to expectation. I analyze this pattern using a two-tailed linear regression on the enclosure data including the medium predator treatment. Enclosure C was excluded for the same reason as mentioned above. The linear regression was significant, however the relationship between treatments was negative, contrary to expectation. Low treatment maintained higher densities of fourth instar predators than medium and high predator treatments. The survival rate from second instar in summer 1992 to fourth instar in spring 1993 was higher in low predator enclosures (26-41%) and in the lake (26%) than 71 01 CO cn 3 o O <3 U TJ CU ,e cj CO C CS 'o . to <T> CU ON -a ^ 05 OH s < s a CO Si C co c 0 c 'u cu si * s CO -t-•PH CO S.S ^ O 01 VI * 2 * -S °*§ +H ~ co « cy -S CN CU 3 H + CO S p. o •a o cs , •si tN ON PH T-H bo w c .£ -3 3 -a bog PJ TJ co S p-o ^ PC> T-H o + U £ ON lH T-H JS 60 .a -p PC co a -a o to 00 <-o a) i—l Tj •£S CO •+-» ^ to 3 o Tj CD (A CU o 73 cu P4 ID O o 00 • pH g o g OJ w P tn O .-s u (0 cu co o O 5 w V w < pJ in CM 00 CO H LO o p + o ^ II H CN II £H o 'co CO QJ PH bO QJ PH pH CC QJ C QJ OH o QJ > co O OH o o ON rH II N CO CO II II , H CN aj c c ca E U p—i E v O PJ © ^ CN O NO A d ^ II II LO E 2^ S lj5u QJ *^ bO cu £ ~ 0 -H ^ QJ OH O QJ > CO O OH CO QJ CO c QJ TJ > P2 cu p* CO TJ QJ PH QJ > QJ PH CO CO PC PC CJ • pH PC u QJ PH CO o PH QJ b0 C TJ 'u X QJ in the medium (11%) and the high predator treatments (6-9%). Enclosure C had the lowest survival rate (3%). Thus, the low predator treatments (excluding C) had survival rates from second to fourth instars similar to those of the lake, and higher than those of enclosures which received larger densities of predators initially (Figure 2.11). In short, predator additions in spring 1992 in the enclosures led to higher densities of fourth instars in the subsequent spring than in the lake, except for enclosure C. Laboratory experiments showed that fourth instar predators have a strong ability to withstand starvation at cold temperature. For predator fourth instar larvae in enclosures, survival over the winter period was not a problem. However, initially high predator treatments maintained lower densities of fourth instar larvae than initially low predator treatments, excluding C. Survival before the winter period but after the second instar recruitment period was a critical point in the persistence of the predator state. 2.3.3 A signal in transition: third instar and fall fourth instar predator dynamics in 1992 The spring predator density gradient imposed experimentally in 1992 sent a disturbance through the community. By comparing the signal sent out and how the predator population responded sequentially through each instar, I can uncover the factors(s) leading to the fourth instar densities in 1993, and the unexpected reversal between low and high treatments. Because overwinter 73 Figure 2.11 Proportion of second instars surviving to fourth instars (summer 1992 to spring 1993) in relation to the density gradient in fourth instar predators in spring 1992. (Survival in Low and High predator treatments are means ± 1 S.D) 50 Survival (%) 40 -I 30 H 20 H io H • Lake O enclosure C (low) ^ Low (Mean, excluding C) A Medium • High (Mean) Spring 1992 predator density gradient (ind. m ~3) 74 survival was not a problem for fourth instar larvae, the fourth instar densities observed in 1993 must have been set earlier in the life history of the predator. I determined where and how the signal changed by graphing the experimental response at different points in the life history of C. americanus (Figure 2.12). Three panels have already been shown in previous figures: the initial predator density gradient (a), the second instar (c) and the spring 1993 fourth instar (f) panels. The first instar information (panel b) is shown for continuity in the data and the picture. First instar densities in enclosures were higher than in the lake. First instar larvae were not expected to exhibit the experimental gradient signal because of the random and donor-controlled nature of the egg laying which masks the survival process of the larvae. I looked at the third instar recruitment densities and the duration of the recruitment period using the same types of tests as in the second instar larvae analysis. Third instar larval densities were expected to reflect the experimental gradient signal as did the second instar with higher maximum densities in enclosures than in the lake and higher maximum densities in the high predator treatments than in the low treatments. This was not the case. In total, 5 out of 8 enclosures had higher third instar maximum densities than in the lake, which is not statistically significant. Only one enclosure recruited substantially higher maximum densities of third instars than in the lake (Figure 2.12, panel d). Between enclosures, there was also no significant difference between the low and high predator treatments. Thus, there was no significant relationship with the 75 co pti a K 5 * <S o C c s Z « c CO cy PC 4J u -s o cy co PH JS rt aa *2 £ fi .3 <u <U QJ -=! ^ .+-> cu 0 ** b -S i; cy CO to •pH PH CN fig 1 S.g> PH CO O r-•H CL C co ST"-* u « C cy o <£ w Ts « 2 ^ " "rt cu >< SJ C8 X -h M)_S co • pH p*H pi tfl h .5 CN rH CN cy PH 60 CO ON ON T-H bi) PH Pl C/JT CN ON ON cti rH o -t-» (Tj a> rH PH fN ON ON H bi) •PH SH P. C/JT PH rS o cn PH P3 •4-1 CA c TJ H CA JH CA _c o u 0) CD CA U • PH PH O •U ni TJ TJ crj PH CU PH CH CJ r o r o e.ui spnpiATpui initial positive predator density gradient. None of the changes in density in enclosures versus the lake and between enclosures were statistically significant (Table 2.7). However, the duration of the recruitment period was longer in all enclosures (8 out of 8) than in the lake (Figure 2.13). In some enclosures, third instars of C. americanus were found late in the fall (Appendix C-3), when most predator larvae were expected to have transformed to their fourth instar; some even survived overwinter. The third instar period seemed to be the time when major changes took place in the initial signal. In the fall 1992, in the new fourth instar cohort, the signal started to differentiate anew. Trends, although not statistically significant (Table 2.8), were apparent. For example, regression slopes were negative, opposite to the original positive predator density gradient. Between the fall and the subsequent spring, densities should have declined through mortality as no reproduction took place. Figure 2.14 represents the relative difference in density between the fall peak in 1992 and the spring peak in 1993. Lake fourth instar densities declined by 39% between the fall and the spring. In enclosures, the response was surprisingly variable. The largest decline, similar to that in the lake, was observed in the high treatment enclosure H (46%). Densities declined by 10 to 25%, in three enclosures, and density declined by only 5% in another one. Meanwhile, densities increased in two enclosures by less than 10%, and by almost 40% in another one. Small declines, and increases of any size in fourth instar density in the enclosures were unexpected as there was no reproduction period for C. americanus in the fall or 77 01 ra o -a o « 0) 45 u • ?H tf) TJ 15 rt TJ fs| 45 rH Oi TJ 45 « ri 45 ^ 15 o> g • r-H Ol tH tH OI g g 3 TJ Oi tf) O tH u oi > n 45 tf> Oi tH tf) o (5 ra ra « H-> tH «) «5 « J2 CU tn 1* O tH &H-2 45 * 4) tf) H U o> 1 TJ (5 ra cu tH tf) o S5 cu cu 45 TJ (5 ra cu 44 ra cu 45 C cu CU tf) 4- g * I tf) 4H cu S5 TJ «U ^ .2 TJ tH ra TJ cu tH Q-, 45 -C • i-H S5 ^ TJ cu (5 bO ra ra is 45 o cj —H cu 45 +•> ~ C tf) 4-> cu cu cu 45 TJ (5 ra cu 44 4H cu 45 (5 cu cu H-» cu 42 TJ _o "tH cu CL, c cu 3 tH u cu tH tH ra tf) CH TJ tH CU 45 CH _o ra tH rJ TJ cu 45 tf) cu tH SP! •H r i -M CJ CH CU CQ a t--C) rH © + rC CN U 2^ ON tH rH £ be .S 45 tf) n S5 MM J^> . „ 'cn bO ci O cu HJ TJ a t. 0 ^ -Q rH 1 + r£J CD U 2^ ON tH rH iS bo « CH .a -c 45 CJ M-l J^> N—' -rH .-r. tf) bO j-O cu J TJ S « •JH Cfl +-> >_| cn H C/3 W H cu TJ r2 •rH ra i cu CH o tf) Ol 15 o TJ 01 00 NO CN X rH CO " + II II M ON II CH cn cu tH 00 su LO o lj CH .2 CU 6 o c • rH CQ cn CH O • rH tf) tf) cu tH bC cu tH tH ra cu S5 cu QH O cu > tf) o CL o o CO o X rH o + rH CN II 00 o 00 II S5 C _o 'cn cn cu tH bo cu tH tH ra cu (5 E t-H E v O rJ CU o cu > tf) o CM x* O o 00 00 a o c •rH CQ *- cn 2 S g P ."S cn oj CJ v 15 O v> rt tH rJ 0 pa m < rJ Figure 2.13 Duration of the main recruitment period for third instar larvae of C. americanus in Shirley Lake and in enclosures. 1992 • 1993 1994 CT3 O C "•4—' CTJ CO _ T3 E CD 3 Q.T3 «s I c o "•4—» CO T3 CD HA E D B C4 Lake -\ • ' 1 i: •! i! i!•:i:•:•:•:i:•!•!•!•:i:•!i!i! i! i! i: i! i :r T—!—s—i—p-T—r"T""r™'r™'t t T T x .y.y.y.y.y.y.y,1 ~T~ 0 12 3 4 Recruitment period duration (Months) ~~r 2 79 cu H-> CO a v. o -a o cs U TJ cu -C cj H-» CO TJ G O to cu PH H-» CU TJ (U to to O M U CU > 2 CN -C ON to cu w co J5 u cu *n HH CO to cu ON <HH CU PH n HH CO *H H4 cu CO cu H CO o PO O <s p« u oo CN CU H CS v CO cu PH CO o u QJ QJ PC TS cS QJ CS '—1 QJ PC QJ QJ > -pJ QJ PQ CO QJ >^ pH -4-» 'CO CO C O QJ TJ G QJ PH ca -4-t CO d PH • PH la TJ PC QJ PH a, .O X HH bO G PC TJ QJ G bO ca c ca X, o CJ QJ QJ PC H-> bO c QJ C •PH QJ •> -t-> CO > H-» QJ QJ H PQ CO S p. -a H § +, PS; CN U £ CJN PH rH JS bfj G -A * pH PH PC co bog O QJ PJ Tj X S J5 •rt co 15 a CD TJ QJ ca ID aj to cu o TJ QJ PH CH co S P. o -a i—1 + u CN ON PH ca H-> ON H CO p-H _d la PC r-+-> PH * >^ o •+-> • PH CO G bO QJ O TJ HJ II CN 00 00 LO d X 9 d ^ II CN <N II ON II 00 LO "ce • i—t s o G • i—i CQ O • PH CO CO QJ pH bO PH PH ca QJ G to G cu TJ co 3 p. O -O o « QJ OH o QJ > co O a, LO o 8 CN r^l T-H d £ CO II G G O • rH CO CO CU PH bO QJ PH PH ca QJ G QJ OH O QJ CD OH V O CO QJ CO G QJ TJ 'QJ > P5H QJ P* ca TJ QJ H-> PH QJ > QJ PH CO ca X, X cj PC u QJ PH CO O G QJ bO G TJ 1J X QJ Figure 2.14 Relative difference in fourth instar Chaoborus density between fall 1992 and the subsequent spring (1993). H E D 1 G B C L (50)% 0% (25)% Relative density difference 25% 50% 81 winter. New fourth instar larvae of C. americanus might have been recruited from the lengthy recruitment period for the third instar larvae, period which unexpectedly encompassed the winter and early spring (see time series for Chaoborus third instar larvae, Appendix C-3). 2.3.4 Result summary Second instar predators in 1992 responded as expected with increased densities in the enclosures (Table 2.9) and the signal recorded was positively related to the predator density signal set experimentally. Increased fourth instar larval density in the spring enhanced summer recruitment of young predators. However, the subsequent third instar response was not related to the initial density gradient. Only recruitment duration showed a difference between the lake and the enclosure community responses. In fall 1992, a weak signal reappeared in fourth instar larva densities but the trend was now inversely related to the experimental gradient. Finally, the fourth instar densities in the subsequent spring showed higher densities than in then lake. The fourth instar predators had the ability to withstand starvation for long periods of time at cold temperature and survived overwinter at high densities. However, low predator treatments maintained higher densities than the high treatments, reversing the expected trend. Thus, some requirements necessary for persistence of the Chaoborus state (Table 2.3) were observed. However, these requirements proved not to be sufficient to allow persistence of the predator state over several predator generations. 82 Table 2.9 Overall results of the impact of the spring fourth instar predator density gradient on the predator dynamics throughout the life cycle Hypothesis: Under reduced Daphnia density and enhanced predator densities, all enclosures are expected to cross the threshold and switch to Chaoborus state Variable PREDICTIONS Lake vs. Enclosures Enclosures: Low vs. high YES/NO Predator Young predator Lake < Enclosures YES Dynamics recruitment density LOW < HIGH YES Fourth instar density Lake < Enclosures YES in spring prior to LOW < HIGH NO recruitment of young instars 83 2.4 DISCUSSION The persistence of the Chaoborus state as an alternative domain of attraction to the Daphnia domain requires first that young predator recruitment in summer be enhanced and second, that fourth instar larvae overwinter well at high densities. Both of these requirements were met in the first year of the experiment. However, these requirements were not sufficient to allow persistence of the predator state over several generations. Developmental delays in the third instar, especially in high density experiments, counteracted the enhanced recruitment of first and second instar predators. The predator density gradient signal was strong enough to travel from one spring to the next, however, in high density predator treatments, the strong signal introduced negative feedback and predator densities were lower than expected. Increased predator density in the spring led to improved conditions for recruitment of second instar in summer. This, in turn, led to higher fourth instar densities in subsequent spring, except in enclosure C, one of the low predator addition treatments. Enclosure C did not recruit fourth instar well but did recruit second instar to a level comparable to that of the other enclosures. Thus, the community crossed the dynamical threshold from the Daphnia domain to the Chaoborus state but the latter did not persist past the fall fourth instar recruitment. 84 The limitation in enclosure C pointed towards the possible location of the threshold between the Daphnia domain and the Chaoborus state. The community in enclosure C was started with the lowest Daphnia reduction (65%) and the second least fourth instar density addition (2.1 times lake level) amongst my experimental enclosures. Based on experiments by Neill (1981b; 1985) in enclosures in Gwendoline Lake, these perturbations were unlikely to yield a switch to the predator state. Neill (1985) used Daphnia reductions (25%, 60%, 90%, 99%) as a method to perturb the zooplankton community in his enclosure experiments in Gwendoline Lake. He reported that a 60% reduction was not sufficient to bring the community to switch states; a 90% reduction was needed. In his predator addition experiments (IX, 2X, 3.5X) (Neill 1981b), he showed that a 2 times lake level addition was not sufficient to change the community to the predator state. Only after additions 3.5 times lake level was the threshold crossed. In enclosure C, values indicating the presence of the threshold in the community dynamics are within a comparable range but tend to be on the low side of the range. Because I combined both types of treatments, the interactions potentially produced a stronger impact than each disturbance alone. On the other hand, the difference in response could also be due to differences in size and depth of the two lakes, or other local conditions. For example, as Shirley Lake is a much smaller and shallower lake (1.2 ha; 10 m max. depth) than Gwendoline (13 ha; 27 m max. depth), it potentially had more nutrient available to the community, which could lead to greater zooplankton productivity. Thus if 85 Shirley Lake zooplankton community sat closer to the threshold, it might have required smaller perturbations to switch states. Gwendoline Lake might require stronger perturbations than Shirley Lake to cross the threshold. Finally, the lack of recruitment of fourth instar larvae in enclosure C in spring 1993 could have arisen in at least two different ways. This enclosure might have responded the least to the experimental disturbance, i.e. although it crossed the threshold, it might have remained close to it, in which case a small disturbance could have reverted the community to the Daphnia domain. On the other hand, the community in enclosure C could have simply dwindled away due to unknown factors, and was on its way to extinction when sampled in spring 1993, rather than displaying Daphnia domain characteristics. I cannot eliminate this explanation especially given that the community did not recruit second instars, not even up to lake level, in early summer 1993 (Appendix C-2). The see-saw in predator densities between the low and the high predator treatments is a strong indication of the presence of the alternative attractor. Low treatment fourth instar densities recruited higher densities in 1993 than their initial predator densities in spring 1992. High treatment densities, on the other hand, recruited lower densities than their initial densities but remained at higher densities than the lake and thus did not return to the Daphnia domain. The pull in opposite directions points to a potential attractor located at a different density level than the Daphnia domain in the lake. 86 In the enclosures (excluding C), survival from second instar in summer 1992 to fourth instar in spring 1993 was inversely proportional to the initial predator density gradient set initially in spring 1992. This density-dependent relationship indicates an upper limit to predator density in the predator state. This key result, combined with information on prey dynamics presented later in Chapter 3, will have important implications concerning persistence of the Chaoborus state The negative impact of initial predator density on the final density of fourth instar larvae did not take place overwinter. Fourth instar larvae are highly resistant to starvation at cold temperature. Cold temperature allowed Chaoborus to survive well overwinter. Bradshaw (1969; 1970) indicated that larvae enter a quiescent phase, a winter diapause, although an "active" one. My own laboratory observations showed that, under winter conditions, the animals were still active. They ate fewer prey than at warmer temperature (pers. observ.), and they developed and pupated at a slower rate (See Chapter 4). Because fourth instar larvae have lower respiration rates in cold temperature (Swift 1976) and greater assimilation rates (Giguere 1980a; Giguere 1981), larvae needed less food and could survive under lower prey density. In Appendix C-4, I showed a dip in larval density in samples from 0-3 m. There was a decrease in October-November and an increase in April-May although no reproduction took place between these dates. When larvae do not need to feed they stay near the bottom of the lake, beyond the reach of the sampling net. This is one explanation for the dip in 87 sampled numbers observed in the winter. Low metabolism and higher assimilation rates in cold water allowed fourth instar larvae to survive with little or no food for several months (Table 2.5 A). Chaoborus can thus survive from late fall to spring even when prey densities are low. The negative density-dependent impact must have taken place prior to entering winter conditions. A long recruitment period relates to improved recruitment conditions for second instars while indicating bad conditions for third instar recruitment. The length of the first instar recruitment period depends on food conditions and on the duration of the egg laying period. Because first and second instar larvae are starvation-prone (Neill and Peacock 1980), the second instar recruitment period can only be increased if first instar have eaten sufficiently to molt into the second instar and the latter also have sufficient food available to survive rather than starve and die. The recruitment period at the third instar is longer for at least two reasons: first, because the second instar recruitment is longer, and second because third instars encounter delays in development in late summer and fall due to prey shortage. A longer second instar period leads to a longer third instar recruitment period. But third instar recruitment period was much longer than produced by the lengthening of the second instar recruitment period. Third instar were found late in the fall, and fourth instar densities were seen to increase or decrease very little overwinter, although natural background mortality should have been present. The transformation of all members of the cohort from third to fourth instars was not yet complete in the fall. Growth and 88 transformation to new instar took place sometime during the winter or early-spring, a clear indication of developmental delays. Such delays, rather than death, are expected in starvation-resistant animals. Thus, the length of the third instar recruitment period was likely due to a combination of a longer second instar recruitment period and third instar developmental delays. I tested the idea of low prey availability for third instar in late summer in Chapter 3. From a theoretical point of view, delays in dynamics can be representative of the vicinity of a boundary between states (Ludwig et al. 1997). The increase in second instar represented the crossing of the threshold between the Daphnia domain and the Chaoborus state. However, the high densities used in the high predation enclosures might have pushed the dynamics of the system close to an upper boundary for the Chaoborus state. This response to 15X increase in predators could be indicative of a system with a small domain and low resilience to disturbance and presents a potential explanation for the fact that the Chaoborus state did not persist over several predator generations. 2.5 CONCLUSION Increasing the predator densities in the spring in the enclosures resulted in higher densities of second instars in all enclosures and in higher densities of fourth instars in most enclosures than in the lake. Higher densities of fourth instar larvae in the spring allowed for better recruitment of the starvation-prone first and second instars in enclosures than in the lake. Fourth instar are 89 starvation-resistant and can survive overwinter at higher densities than in the lake. However, developmental delays indicated that limitation in the system occurred before the winter period. The bottleneck, previously experienced by the first and second instar larvae, moved to the third instar larvae. Enhanced recruitment of young instars and good overwinter survival are necessary but not sufficient elements to guarantee persistence of the predation state over several predator generations. Next, I examine the prey dynamics in the light of the predator responses above to understand the lack of persistence of the Chaoborus state under the experimental conditions. 90 CHAPTER 3 PREY DYNAMICS, SHORT TIME SCALES, AND PREDATOR RECRUITMENT 3.1 INTRODUCTION Dynamics of Daphnia and solitary rotifers, the prey populations, play an important and intricate role in the community switch from the Daphnia domain to the Chaoborus state, and in the potential for persistence of the latter. The switch between states depends on a rerouting of the food resource for the prey from the Daphnia population to the solitary rotifers, the small prey. This resource redirection must take place prior to and during the recruitment period of the young predator instars. On the other hand, the persistence of the Chaoborus state depends on the feedback of the predator dynamics on the prey assemblage. High predator densities, expected in the Chaoborus state, require sustaining high prey production, from one predator generation to the next. In this chapter, I test that increased predator densities in the spring reduces the competitive impact of the Daphnia population on the solitary rotifers prior to young predator recruitment (see Table 1.1). I also test the impact of high predator densities on the prey dynamics throughout the predator life cycle and on the persistence of the Chaoborus state. Specific predictions related to the enclosure experiments are presented in section 3.2.3. 91 Redirection of the food resource to small prey depends on the ability of Daphnia to sequester resources, ability which can be reduced or enhanced by different factors such as temperature and predation. Daphnia has a big potential for food consumption and can reduce resources below the level required by other zooplankton (Lampert et al. 1986, cited in Jiirgens 1994a; Glide 1988) such as small cladocerans (e.g. Daphnia vs. Bosmina: Jiirgens et al. 1994b), copepods and rotifers (Gilbert and Stemberger 1985; Gilbert 1988a). Daphnia can reproduce quickly and have a major impact on phytoplankton resources as demonstrated by the spring clear water phase in many lakes where reduction in phytoplankton density by Daphnia feeding produces high water transparency (e.g., Edmonson and Litt 1982; Lampert et al. 1986; Rudstam et al. 1993; Jiirgens 1994a). In lakes where Daphnia is low or absent (Havens 1990; Stenson 1990), other species, such as rotifers, or Bosmina, are abundant throughout the season instead of being limited to early spring and/or fall peaks. Daphnia is a key player in determining zooplankton community structure. Water temperature affects metabolism and thus consumption rate. At low temperatures, Daphnia reproductive, developmental and growth rates decline (Hebert 1978; Neill 1981a; Orcutt and Porter 1983; Berberovic et al. 1990). These declines in physiological rates reduce food requirement for survival and growth. Daphnia rosea is a cold water species (Neill 1981a; Walters et al. 1987), compared to species such as D. pulex. Cold temperature is even less of an impediment for 92 many rotifers. They are present and reproduce before Daphnia emerge from resting eggs in the spring and are also present and reproduce in the fall after Daphnia, having produced resting eggs, disappear from the water column. Thus, in cold temperature, the competitive influence of Daphnia is reduced or absent altogether. Cool years could facilitate the switch from the Daphnia domain to the Chaoborus state. The impact of the predator on Daphnia resource consumption is indirect. Predators reduce Daphnia population density and modify its population size structure which then affects Daphnia per capita feeding and reproductive rates. Predators such as fish deplete the larger size Daphnia (Brooks and Dodson 1965; Gliwicz 1985). This reduces both the reproductive output and the per capita feeding rate as larger Daphnia are generally the reproductive individuals in the population and they can clear a larger volume of water per unit time than smaller individuals (Borsheim and Andersen 1987; Peterson et al. 1987). Fish is absent from Shirley Lake. Instead, invertebrate predators, such as Chaoborus, feed on small size classes (Fedorenko 1975b; Neill 1981b; Buns and Ratte 1991) compared to fish and reduce Daphnia recruitment to the reproductive population. Predation can reduce resource consumption by Daphnia populations. In my enclosure experiments, started in 1992, reduction of Daphnia density and additions of Chaoborus were combined to study the impacts on both the Daphnia 93 population and the solitary rotifer populations. In 1993, fourth instar predator densities, recruited from the previous season, and cool spring temperature combined to delay the onset and slow down the increase in Daphnia abundance which in turn was expected to affect the solitary rotifer population dynamics. Both the strength of the disturbance and its timing play a large role in the resulting dynamics and in determining presence and persistence of alternative states for the zooplankton community. 3.2 MATERIALS AND METHODS 3.2.1 Field experiments Details for the setup and design of the field enclosure experiments in Shirley Lake are described in Chapter 2 (section 2.2.1 to 2.2.3). Samples for zooplankton and rotifer in enclosures and the lake were collected on the same schedule as for the predators (for details, see Chapter 2, section 2.2.3). Generally, predator daytime samples, large zooplankton samples, rotifer samples, and the water temperature data were collected between 10:00 and 14:00 Pacific Daylight Saving time. I collected predator and large zooplankton samples by vertically hauling a 0.4 m mouth diameter Wisconsin type net, 102^im mesh size, from a depth of 3 m. At the deep lake station 2, a 6 m vertical haul was also collected. The rotifer samples were collected in the same manner but with a 0.3 m mouth diameter Wisconsin 94 net, with a 50 urn mesh size. I preserved all samples in glass jars using 5% sugared formaldehyde solution. A three-year time series was collated from monthly samples for 1992 to 1994 for total Daphnia and total solitary rotifer densities for each enclosure and for the lake station (Appendices D and E). The solitary rotifer data represented total densities for non-colonial rotifer genera Keratella, Kellicottia and Polyarthra, genera which are potential prey for all Chaoborus instars and necessary prey for younger instar larvae (Moore and Gilbert 1987; 1988a; 1988b). Solitary rotifer time series for 1992 to 1994 were erratic and difficult to interpret (Appendix E). The monthly sample interval did not provide the required details to test the predictions presented in the next section. The shorter interval samples (semi-weekly, weekly and bi-weekly) provided better information which is presented later in the Results section. Daphnia time series showed that Daphnia was present seasonally in the lake throughout the study period from 1992 to 1994 (Appendix D). Most enclosures recruited Daphnia seasonally throughout the study period. However, no substantial densities of Daphnia were found in enclosure G from the onset, and in enclosure C after August 1992. Daphnia recruitment failure was evident in enclosure G where peak monthly density only reached about 30 Daphnia m" (10- to 200-fold lower than densities reached in the other enclosures and the lake), and where the population disappeared from the water column in September, rather than later in the fall. 95 Water temperature at 1 m interval was sampled using a Par battery operated bilge pump and a thermometer (±1°C). Mean monthly water column temperatures were calculated using the recorded temperature profiles from the surface down to 7 m for each date within a month. I also calculated the mean seasonal water temperature using the monthly mean water column temperatures for dates between May and October inclusively. 3.2.2 Identification and counts I used Edmonson (1959) as a reference for identification. Macrozooplankton (e.g. cladocerans, copepods) were identified to species under a Wild M5 stereoscope at 40X. The rotifers were identified to genera, under a Nikon inverted microscope at 100X. I counted macrozooplankton and rotifers under a Wild M5 stereoscope at 12X, 25X or 40X depending on zooplankton size and sample density. Chaoborus were identified to species, using Saether (1970) and Borkent (1979), and identified to instars using Fedorenko and Swift (1972), under a stereoscope at 25X, and at 40X for younger instars. Counts were done in grided petri dishes (Edmonson 1971b), under a Wild M5 stereoscope for all animal categories. Samples estimated to contain less than 400 individuals of the most abundant species in the sample, were counted in their entirety. Samples estimated to contain greater than 400 individuals were split, and one or two subsamples were counted, as required, to count about 100 individuals of the most abundant species in the sample (Edmonson 1971b). 96 Chaoborus third and fourth instar larvae were counted at 12X in entire samples, as their elongated shapes prevented random sample splitting. For all other categories, dense samples were split using a 250 ml Folsom splitting wheel with 4 divisions (4 X 1/4 subsamples) (Edmonson 1971b). When necessary dense subsamples were resplit and the volume to count was reduced to 1/16 or 1/64. Here again, the subsamples to resplit and to count were chosen randomly. 3.2.3 Predictions for prey dynamics based on experimental design Based on the hypothesis that the zooplankton community in Shirley Lake can function in two different domains of attraction and based on the current experimental design of Daphnia reduction, directly and indirectly by predator additions, I expected that the zooplankton community in enclosures should switch from the Daphnia domain, as present in the lake, to the Chaoborus state (see Section 2.2, for details of community functioning). The switch to the predator state was expected to reduce Daphnia densities and increase solitary rotifer densities prior to the recruitment of second instar predators (Table 3.1). Enhanced Chaoborus fourth instar densities in spring were expected to reduce and maintain lower Daphnia densities in the enclosures than in the lake prior to the young predator recruitment period. Amongst enclosures, low treatments were expected to yield higher Daphnia densities than the high predator treatments. 97 In response to these low Daphnia densities, I expected that the solitary rotifer populations would increase during the same time period. Thus, rotifer densities should be higher in the enclosures than in the lake, and higher in high predator treatments than in low treatments (see Table 3.1). Persistence of the Chaoborus state requires that high densities of predators survive overwinter and that they affect negatively the Daphnia population in the spring. The predators have short-term impact on the Daphnia population by delaying the onset of the population and/or longer-term impacts by affecting population growth rates. Thus, in enclosures which recruited fourth instar predators in higher densities than in the lake, I expected that Daphnia population onset would take place later, and/or that Daphnia densities would be lower for a longer period of time than onset time and densities in the lake. I addressed these points by using the date of appearance of Daphnia in the samples. A potential problem is that Daphnia could have been present in the lake before I detected it. Presence of Daphnia was detected in my samples when their density in the water column exceeded 2-3 Daphnia m" . A second problem is that, in 1992, the enclosures were started when Daphnia was already present in the lake, thus I could not use this information to study the influence of the density of predators on the onset of the Daphnia population. Instead, I used data from spring 1993 where predators were present at different density levels in the enclosures prior to 98 Table 3.1 Predictions for the relationships in prey densities between the lake and the enclosures, and between the low and high treatments in relation to the experimentally-imposed predator gradient. Hypothesis: Under reduced Daphnia density and enhanced predator densities, all enclosures are expected to cross the threshold and switch to Chaoborus state Variable PREDICTIONS Lake vs. Enclosures Enclosures: Low vs. high Correlation with spring predator gradient Daphnia Daphnia density prior to recruitment of young instars Lake > Enclosures LOW > HIGH -Daphnia appearance date (Julian day) in samples Lake < Enclosures Solitary Rotifers Solitary rotifer density prior to recruitment of young instars Lake < Enclosures LOW < HIGH + 99 Daphnia coming out of resting eggs. Thus, I could study their impact on the beginning of the Daphnia population growth. I expected that Daphnia would appear sooner in the lake than in the enclosures with high fourth instar recruit densities (Table 3.1). Delays in the onset of Daphnia population and lower Daphnia densities compared to the lake prior to recruitment of young predator instars would support the idea that feedback loops with the potential to generate alternative states were present. 100 3.3 RESULTS 3.3.1 General trends in prey population dynamics The full time series (1992-1994) are presented in Appendices D and E for Daphnia and rotifer densities respectively. Figure 3.1 represents Daphnia densities on the date when first instar predators appeared in the samples (indicated by inverted triangles in Appendix D) in relation to density of fourth instar predators earlier in the spring for each year of the study. Over the three-year monitoring period, the Daphnia population in the lake reached at least 1000 individuals m"3 at the time when first instar predators appeared in the samples while, in the enclosures, densities were generally lower than 1000 Daphnia m" (Figure 3.1). In enclosures, such densities were reached only later during the predator second instar, if at all (Appendix D). The solitary rotifer densities at the time when the first instars appear each year showed no specific relationship between solitary rotifers and fourth instar densities in lake and enclosure samples (Figure 3.2). Interpretable patterns related to predictions from Table 3.1 appeared only when looking at data on the time scale shorter than one month and are addressed below. The 1992 time series presents bi-weekly to monthly samples for Daphnia (Figure 3.3) and for the solitary rotifers (Figure 3.4). The inverted triangles represent the date of appearance of the different predator instars in the lake samples. First instars appeared in enclosures on the same date as in the lake. However, date of 101 Figure 3.1 Daphnia density on the date when first instar predators appeared in Shirley Lake and in the enclosures each year in relation to the density of fourth instar predators in each spring. (Appearance date in monthly samples: June 30,1992; June22,1993; July 5,1994) Predator addition treatments Lake Low Medium High 104 Daphnia density 2 (ind. m"3) 10 " 1023J 101J 101 [94 10 J -i • • • • • L92 L94| L93^ E93 B94 • B92 • J92 C92^ * H94 E92 J93 • ^2 H92 D92 D93 J94 • • E94 H93 B93 • • • . 10' 10J Fourth instar predator density in spring (ind. m-3) 102 Figure 3.2 Density of solitary rotifers on the datewhen first instar predators appeared in Shirley Lake and in the enclosures each year in relation to the density of fourth instar predators each spring. (Appearance date in monthly samples: June 30,1992; June22,1993; July 5,1994) Predator addition treatments Lake • Low A Medium • High Solitary rotifer density (ind. m "3) Fourth instar predator density in spring (ind. m"3) 103 Figure 3.3 Time series for total Daphnia density May to October 1992 following a variable sampling interval. (a) in medium (I) and high (D, E, H) predator treatment enclosures (b) in Shirley Lake (L) and low (C, J, B, G) predator treatment enclosures (Inverted triangles represent onset of instar recruitment in the lake) Vertical guidelines: Julian day 154 : fourth instar predators have been reduced through pupation after Julian day 168 : appearance of second instar larvae in enclosures Julian day 266 : Daphnia population decline due to environmental factors Horizontal guidelines: 1000 Daphnia m "3 Julian Days 104 Figure 3.4 Time series for total Solitary rotifer density May to October 1992 following a variable sampling interval. (a) in medium (I) and high (D, E, H) predator treatment enclosures (b) in Shirley Lake (L) and low (C, J, B, G) predator treatment enclosures (Inverted triangles represent onset of instar recruitment in the lake) Vertical guidelines: Julian days 154 to 168: period where predation was low and where the relative strength of competition by the Daphnia population on the rotifers can be tested Horizontal guidelines: 10 000 solitary rotifers m"3 Julian Days 105 appearance of second, third and fourth instars in enclosure samples could take place earlier or later than in the lake samples. The lake and low predator treatment enclosures J and B followed a similar seasonal pattern with the low predator enclosures showing lower Daphnia densities than the lake till well into the second instar recruitment period. On the other hand, the medium and high predator enclosures showed low Daphnia densities (below 1000 individuals m"3) throughout the summer and did not reach a plateau before the fall population decline. Solitary rotifer densities also showed patterns that were strikingly different for the medium and high predator treatments (Figure 3.4, panel A) than for the low treatments and the lake (Figure 3.4, panel B). In the latter two categories, solitary rotifer densities oscillated around 10 000 individuals m" with rotifer densities above 10 000 individuals m" at the end of the second instar and the beginning of the third instar recruitment periods. In the high predator treatments, rotifer densities showed a downward trend from spring to September, when they showed an increasing trend till the end of the ice-free season. This means that contrary to the lake and low predator treatments, high predator treatments showed a dip in rotifer densities below 10 000 individuals m"3 at the end of the second instar and the beginning of the third instar recruitment period. In the following sections I use data from 1992 to demonstrate the impact of the predator gradient on the prey populations prior to young predator recruitment 106 and data from 1993 to highlight the impact of the predator on Daphnia populations early in the spring. 3.3.2 Daphnia population dynamics: influence of spring predator density and temperature 3.3.2.1 Impact of the 1992 spring predator density gradient on Daphnia densities The initial impact of the experimental treatments (Daphnia reduction and predator additions) was to decrease the Daphnia densities in May 1992 in the enclosures compared to the lake (Figure 3.5). All the enclosures are functional in may 1992 and are represented on figure 3.5. However, enclosures A, F and G failed early in the experiment. I excluded these enclosures from this and subsequent analyses. Here, I analyzed the data using a one-tailed linear regression. The relationship between the logarithm of Daphnia density versus the logarithm of fourth instar density initially present in the enclosures was negative (p=0.008, R =0.64; one-tailed test; df: (1,6)) as expected from predation theory. Thus, the predator density gradient applied experimentally in the spring resulted in a Daphnia density gradient. Higher densities of fourth instar predators in the spring introduced delays in the increase phase of Daphnia population in enclosures. Larger delays in Daphnia increase represent larger windows of time for young predators to recruit. I used the date when first instar predator appeared in the samples as day 0. Using 107 Figure 3.5 Initial Daphnia densities in the lake and in the enclosures after predator additions (May 26th, 1992). Regression line for all data points except for those between parenthesis (G, F, A) where Daphnia or the predators failed to recruit early in the experiments Log (Daphnia +1) = 5.1 -1.5 Log(predator +1) R2 = 0.642 Predator addition treatments High 101 102 103 Spring 1992 Fourth instar predator density gradient (ind. m~3) 108 Daphnia density in the lake at day 0, I calculated the delays for Daphnia populations in enclosures to reach equivalent or higher densities. The low predator treatment enclosures (B, J) required four weeks in 1992 and 8 weeks in 1993 (Table 3.2) to reach Daphnia densities equivalent to those in the lake when first instar predators appeared (see Figure 3.1). The medium (I) and high (D, E, H) predator treatment enclosures required at least twelve weeks to reach those same Daphnia density, except for enclosure E in 1993 which reached high densities of Daphnia at the same time as in the lake. Higher densities of fourth instar predators in spring 1992 resulted in lower Daphnia densities prior to second instar recruitment (June) in the enclosures than in the lake (Figure 3.6). I selected the time period after most fourth instars had pupated, but before the main recruitment period of young predators (from Julian day 154 to 168) (Figure 3.3). This period included three daytime sampling dates (June 2nd, 9th and 16th, 1992). For the lake station and each enclosure, except G which did not recruit substantial Daphnia, I plotted the average density over this two-week period (Figure 3.6). All the enclosures recruited lower density than the lake. A one-tailed binomial test (Table 3.3) was statistically significant. The linear regression (Table 3.3) was significant and the slope was negative as predicted in Table 3.1. Amongst enclosures, the predictions were also upheld with higher Daphnia densities in low predator treatment than in high treatments. The results of a 109 Table 3.2 Impact of fourth instar predator density gradient on Daphnia population increase: Delays (in weeks) in enclosure Daphnia population in reaching densities similar to those found in the lake at the time when first instars appeared. Treatment Station 1992 delay (weeks) 1993 delay (weeks) High H >12 12 E >12 0 D 12 12 Medium I >12 8 Low B 4 8 J 4 >8 Lake L — — 110 Figure 3.6 Mean density of Daphnia (for Julian days 154 to 168, June 1992) in relation to the experimental gradient in fourth instar predator in spring 1992. Log(Daphnia +1) = 4.5 - 0.93 * Log(predator +1) R2 = 0.785 Predator addition treatments Lake • Low (excl. G) Medium High June 1992 Daphnia density (ind. m'3) 1q3 , Spring 1992 Fourth instar predator density gradient (ind. m "^) 111 •w (A CO o o « 3 cu cu CU -o TJ 3 « EA CU 1H CO O CN ^-3 s cu 45 TJ CU 45 u TJ 3 TJ "o co CU tH H-> cu 45 H-> TJ CU CA CA o tH TJ 3 rt cu 4<, rt cu H-» 3 cu cu cu CM ON ON CU 3 3 cu & s *S to CA 3 CU CU cu I-i S3 TJ S 3 CA CA is O a -s 3 CU Q 13 Q o ** .3 « CA m TJ ^ S ^ m CU 4H OH MH —H MH 60 £ ° .S 5 H-> CA CA m CU CU = H H S CO CO CU •8 g + O CN "S ON |S U bO 3 CM tH CA CO .S 3 •i-H tH --H 3 CA o c MH O) *— TJ bC O rJ -4-> • '55 CN (5 °^ S ON TJ 1-1 CU « c: s 3 Q bO O ra u CO cn 3, o CO cu TJ 4H '3 4-> I CU w H CO cu 3 o TJ CU (H CH x* o o 00 00 .2 B o 3 • »H PO CO cn O H-> NC *CA m 3 cu A TJ W <S s < 4K HH a. « •i-H CM O o X 8 LO 9 ^ LO o II CM NO 3 _g 'CA CO cu tH bO cu tH tH CO CU 3 cu CM O cu > CO o CM CN| o o © 00 CN N ON CN 3 cu 3 -M • rH 45 I 3 3 cs o l-H A o o © X 00 LX o I CO ^t5 >" PH o lx NO LO 3 O • -H CA CO OJ tH bO cu tH tH cd 3 cu CM O cu > CO O CM linear regression showed significant and negative relationship (Table 3.3). Thus, more fourth instar predator in the spring resulted in lower Daphnia densities prior to recruitment of the second instar predators. Increased densities of fourth instar Chaoborus in the spring delayed Daphnia population increase in early summer, and provided lower Daphnia densities at the time when young predators started recruiting. Higher densities of predators enlarged the delay in reaching high densities of Daphnia. 3.3.2.2 Daphnia population onset: influence of water temperature and predator density in the spring The 1992 data presented above addressed the impact of the spring fourth instar predators on Daphnia densities prior to second predator instar recruitment. To address the impact of the predators on the Daphnia population in early spring, I needed to use 1993 data because enclosure experiments in 1992 were started after Daphnia had already appeared in the lake. In 1993, predators overwintered in the enclosures and were present in high densities in most enclosures before Daphnia hatched out of resting eggs. I thus could observe the impact of naturally recruited, but different densities of fourth instar predators on the onset of Daphnia population dynamics early in the spring. Daphnia could have been present in the lake before I could detect them in the samples. Daphnia population onset was defined as the Daphnia population 113 reaching a minimum level of 1 Daphnia per sample, i.e. 2-3 Daphnia m'3. I expected that the predator could delay the start of Daphnia population increase by predating either on Daphnia hatching out of resting eggs or on their offspring. Delays in the onset of Daphnia population increase, and differences in timing and densities of Daphnia population peaks occurred between years, and between enclosures (Appendix D). Because 1993 was a colder year than 1992, and because cold temperature can slow invertebrate metabolism, I first determined the impact of temperature on the time of appearance on Daphnia in the spring. To evaluate the impact of temperature on the onset of Daphnia population, I used as day zero the first day when Daphnia appeared in the lake samples in the spring of 1992. I calculated the difference in Julian days in the onset of Daphnia population. I graphed the delay in the date of appearance of Daphnia in the lake samples between years in relation to mean spring temperature of the water column, and in relation to densities of fourth instar Chaoborus larvae in the spring (Figure 3.7). The longest delay in Daphnia appearance in the lake between years was 19 days (Figure 3.7 A), and occurred in 1993, which had the coldest spring and was the coldest year overall in the study. There was no clear relationship between the length of the delay in Daphnia population appearance and the fourth instar larval densities over the range observed in the lake in the spring (Figure 3.7 B). The yearly mean temperature of the water column related well to Daphnia maximum and yearly mean density (Figure 3.8). 114 Figure 3.7 Delay in Daphnia population onset in 1992 to 1994 in relation to A) mean water temperature in Shirley Lake in springtime (mid-May to mid-June) B) predator density in Shirley Lake in spring 1992 to 1994 A) CO D 01 (0 G O o 9.5 Temperature (°C) 3 PH o OH <3 K -s: a. cs Q d • r-l D B) Predator density (Chaoborus fourth instar larvae m 115 Figure 3.8 Maximum and yearly mean Daphnia density in 1992 to 1994 in relation to yearly mean temperature. Dotted lines below the temperature axis represent ± 1 s.e. of the yearly mean temperature for each year. -9 Maximum • Yearly mean Daphnia density (ind. m "3) 6000 5000 H 4000 H 3000 H 2000 H 1000 Temperature (°C) 116 Using the same approach, I examined the influence of spring predator density in the enclosures, and in the lake, on the timing of the onset of the Daphnia population. The longest total delays were 75 days and occurred for two enclosures (B and H) with densities of Chaoborus greater than 155 predators m" combined with the cold spring temperature of 1993. I used the total delay obtained for each enclosure and subtracted the delay associated with temperature as observed in Figure 3.7 (e.g., 19 days for 1993 data) to obtain the delay associated with the predator effect. I graphed the predator-induced delay in relation to spring predator density after removing the effect of temperature in Figure 3.9. Enclosures with delays equal to zero had delays indistinguishable from temperature delays alone. Stations B93 and H93 showed delays almost four times greater than the delay observed in lake (L93) for that year. Predator densities higher than 155 fourth instar larvae m" in the spring delayed the onset of Daphnia population in the enclosures. An important predation pressure was removed from Daphnia population after fourth instar predators pupated and metamorphosed in early June 1992. However, Daphnia densities in enclosures did not recover to levels equivalent to that of the lake by the time the third instar predator reached their maximum recruitment in August (Figure 3.10). High predator treatments maintained lower Daphnia densities than the low treatments (Table 3.4ii). 117 Figure 3.9 Delay in Daphnia onset in the spring in relation to the density of fourth instar predators in springtime of each year. Predator addition treatments Lake • Low A Medium High -3 60 A D cy CO G o el *•*•> OH O OH <S • w s »« a. Q c «a 'cy D ro Oi co 40 TT PO —I —i 20 0 l l • l l 200 250 Fourth instar predator density in springtime (ind. m "3) 118 Figure 3.10 Daphnia densities in August 1992, when third instar predators reached their maximum density, in relation to the density gradient in fourth instar predator in spring 1992. Regression line for all data points except (C) where Daphnia recruitment had failed by this date Log (Daphnia +1) = 5.1 - 0.95 Log(predator +1) R2 = 0.815 Predator addition treatments High 101 102 103 Spring 1992 fourth instar predator density gradient (ind. m 119 H-I TJ to " to to" 3 01 V u O 3 -a to 0 o s« vj oi 0 01 43 TJ ** O" Tj H-l >U CO ft) TJ C 1 8 J3 ^ H-l , ^ 01 CN J3 °^ •is ON Oj w to to rv to M -3 -H ft) to > OI ?! c w 3 TJ o to « 0 *a « « J H 5 a. * <" « OH x c go 'j? 01 C c OHMH O _^TJ ° 43 w oi 00 S " 01 0) o e • lH H-l H-l to to ft, «• tU U ft, H H 42 Oi 43 rt H i. + U bo c3 n T—I ^ + Tj bh o c MH QJ "~^Tj bO O II X PK bO O t-J II OH 8 « •rt co H_» -H 3 o CO 0» H-l TJ H-l I QJ w ° w H to cy o TJ Oi -H CD O O o ON LO 00 d o i LO II CN LO (3 o • .—t to to 0) -H bo QJ lH -H CS QJ c QJ OH cn o cn .2 CS bO o U •a « (3 A •S3 a.^ « Q CM O CN CO LX N CO CO II II i-H CN J3 S3 QJ s3 cs O r—I E A o LO CN O X VO d tx CO CO O II (N >H 04 C _o 'co co QJ SH bO QJ CS QJ 53 QJ OH TD CD > H-» CS bO QJ S3 3.3.3 Relationship between Daphnia population and solitary rotifer population dynamics The spring predator density gradient generated a Daphnia density gradient in the enclosures which was maintained at least till the recruitment period of second instar predators, and longer in some enclosures. I expected solitary rotifer densities to vary inversely in proportion with the Daphnia gradient (Table 3.1). To evaluate rotifer dynamics in relation to the Daphnia density gradient, I focused on a period, prior to second instar recruitment, where both competition and predation on the rotifers should have been limited. I used the period including Julian day 154 to 168, a fourteen day period with three sampling dates (Figure 3.4). The inverted triangles on this figure represent timing of appearance of the predator instars in the lake. Previous to day 154, rotifer densities were declining in the lake and most enclosures. In the lake, the decline may have been associated with the increase in Daphnia density. In the enclosures, Daphnia densities were low (generally fewer than 300 individuals m"3; Figure 3.3) and rotifer densities were expected to increase. However, fourth instar predator larvae added to the enclosures may have exerted sufficient predation pressure to exceed effects of the competitive release on the rotifer population. Fourth instar predators can feed on rotifers (Moore and Gilbert 1987; 1988b; 1994), especially when other prey such as Daphnia are rare. As the predators entered the pupation period, their density declined rapidly, thus lowering the predation pressure on the rotifer populations. After day 168, second instars started to recruit in larger 121 densities in most enclosures, bringing a sharp increase in predation pressure. Thus, between day 154 and 168 on figure 3.4, rotifer populations were expected to be under reduced predation pressure from Chaoborus, and rotifer populations are free to vary in relation to the strength of the competitive pressure from Daphnia. Based on competition theory, I would expect a negative relationship with more rotifers present as Daphnia densities were lowered. Thus the lowest rotifer densities were expected in the Daphnia domain, and the higher rotifer densities in the Chaoborus state, where competition was hypothetically reduced by predation on Daphnia, the dominant competitor. For each station, I averaged the density of both Daphnia and of solitary rotifers over the three sampling dates (Figure 3.11). The relationship between Daphnia and rotifer densities showed a negative trend but was not significant (Table 3.5 i). The relationship is not linear. A three-fold decrease in Daphnia densities in the lake compared to the enclosures, resulted in a five-fold increase in rotifer densities (Figure 3.11 A). A linear regression on the data was not significant but a second order polynomial model fit relatively well (Table 3.5 i). Amongst enclosures alone, a fourteen-fold decrease in Daphnia densities (700 - 50 individuals m" ) from low to high predator experiments resulted in a two-fold decrease in rotifers (33 000 - 18 000 individuals m"3) (Figure 3.11 B). Both a linear 122 Figure 3.11 Mean density of solitary rotifers in relation to Daphnia mean densities for Julian days 154 to 168 (June 1992). A) Lake and enclosures: second order polynomial relationship. B) Enclosures only: linear and second order polynomial relationships. Regression equations are presented in Table 3.6 Predator addition treatments • Lake * Low A Medium • High V CO -t-> ; O T3 5-1 O CD A) B) IC 10' Daphnia density (ind. m~3) 123 41 43 .. +* ii e CO > CO £ S cu °£ »© Tj o cl <S 42 U o ca CO cu PH _ CO TJ O cu —i * % CJ *H S 43 TJ TJ s S $ « 3 « TJ cu .-a 2 a d -C ^ cu £ « TJ _H -S CS — S « cu cu -5 5 I It 's JQ g - d H .H ! U CO 2 pH "PH CU « "2 gj CU « ca TJ 4, CO 4H C cu -3 "H a o co 5 *H cu -"Si x s S — .cl C! co m O 4, c ca X V TJ ° ii S 43 £ 5 ?1 •PH ca CO CO > ft) ft, > h HP5 CO cu H CO C cu TJ CS e PS: ' PX. Q bO O HJ II X CM ON ON PH cu PH 3 + H^^ •13 ON co ON S3 rH CU TJ pH cu o • bO O PJ II CO H-» CO QJ PH "ca CJ <a cn 3 O TJ P^H •PH ca -pJ cu I-, c W H CO cu S3 o TJ CU PH OH "tf O © o 00 00 T& •PH s o g LO 00 X 00 rH d I oo "tf X ON d X o LO rH ^ CO ON + rH 00 CM O d d II 11 I'I r-J II CH* P4 >H NO S3 _o 'co co QJ PH bC QJ PH PH ca QJ S3 in o d V LO x la 6 o c pb 'o PH QJ OH O ircn •5 O S w a v H-> p pq fr< •rH y V r-H -I-H o CO QJ > co O OH TJ c QJ PH PH ca cu $3 • i—I i-H S3 O CO 00 o NO 00 CO N CN CM PH CM S3 S3 QJ S3 P2 <-i S3 S3 ca co o o X "tf ^ rx o o + 00 ON O II CM LO S3 o •pH CO co QJ PH bO QJ pH PH ca QJ S3 X U HH X v O rJ QJ OH O QJ > CO O OH LO o d V LO ON d II CN X co d i X NO rH + "tf CN II >* "tf ca a o S3 pb IS PH TJ £3 QJ pH -pH PH ca QJ S3 o 2 and a second order polynomial model fit the data, and the nonlinear pattern has a higher coefficient of determination (Table 3.5 ii). The observed pattern amongst enclosures, more rotifers when more Daphnia were present, was contrary to expectations based on competition theory. This pattern might be related to predator recruitment. Later in the summer, when third instar predators were recruiting, both Daphnia and solitary rotifer densities generally decreased with increased predator densities in the community in the spring (Figure 3.12, panel c and f). The negative trend in Daphnia relationship to the spring predator gradient did not change qualitatively between June and August, except for a gradual increase in density in the lake and in most enclosures through the summer (Figure 3.12, panels b and c). In the case of the solitary rotifers in May, there is no clear relationship with the spring predator gradient (Figure 3.12, panel d). Moreover, the qualitative relationship with the predator gradient was altered dramatically between June and August (Figure 3.12, panels e and f). In the lake, rotifer densities increased slightly from June to August while in most enclosures there was a drastic decline (1 to 2 orders of magnitude). The lake had higher densities of both Daphnia and rotifers than most enclosures (Figure 3.12, panels c and f). This did not follow the predicted trend. Finally, I address the response of the prey populations in 1993 and 1994 after the fourth instar predators had disappeared from the water column but where, 125 Figure 3.12 Solitary rotifer densities and Daphnia densities in 1992, initially in May, during young predator recruitment in June, and during third instar recruitment in August in relation to density gradient in fourth instar predators in spring 1992. Prey dynamics in 1992 Predator addition treatments Lake Low Medium High May June Aug Daphnia density (ind. m "3) 1 01 10° Solitary Rotifers density (ind. m ~3) 1 05 104 1 03 102 1 1 d) i 1 ' e) 1 r f) • • • AH • • • - • 1 o5 104 1 o3 102 1 01 1 02 1 03 1 01 1 02 1 03 1 01 1 02 1 o3 Spring 1992 Fourth instar predator density gradient (ind. m "3) 126 contrary to the case discussed previously for day 154 to 168 in 1992, a new predator cohort has failed to recruit. Based on their competitive interaction and the lack of predation by Chaoborus, I expected a negative relationship between maximum Daphnia density and solitary rotifer densities. Enclosures B (low), I (medium) and D (high predator treatment) lost their predator cohort in 1993 while enclosures J (low), E and H (high) did so in 1994. In 1993, only enclosure I, compared to the lake, followed the expected trend (Figure 3.13). Enclosures B and D had lower densities of both Daphnia and solitary rotifers compared to the lake. In 1994, enclosure J and E recruited low densities of solitary rotifers considering that their Daphnia densities were 10 to 25 times lower than in enclosure H and in the lake (Figure 3.13). Enclosures, such as B, D, E, J, with low densities of both Daphnia and solitary rotifers after the predator impact had weakened had most likely become dysfunctional. Populations of Daphnia and rotifers in enclosure I in 1993 and in 1994 (see Appendices D and E) seemed to vary inversely from each other. In enclosure H, rotifers increased as Daphnia density decreased at the end of the 1994 season. In summary (Table 3.6), Daphnia densities in 1992 responded as expected to the spring predator density gradient. The negative relationship was maintained through the season till the Daphnia population declined in the fall. In response to reduced Daphnia densities in enclosures, solitary rotifer densities in 1992 increased and reached higher levels than in the lake prior to young predator recruitment. However, in the high predator treatments, solitary rotifers in 127 Figure 3.13 Solitary rotifer densities in relation to Daphnia maximum densities after enclosures have lost their predator cohort A) for enclosures that failed in 1993 B) for enclosures that failed in 1994 ro 13 & •rH CO C 0) 13 u 0) o SH O cn 10 2 103 i • . . . 104 . . .. 1 105-: io4-• L 103-: • B A I 102 -i • D 101 ' Daphnia maximum density (ind. m "3) 128 Table 3.6 Overall results of the impact of the 1992 spring density gradient in fourth instar predators on the prey dynamics Hypothesis: Under reduced Daphnia density and enhanced predator densities, all enclosures are expected to cross the threshold and switch to Chaoborus state Variable PREDICTIONS Lake vs. Enclosures Enclosures: Low vs. high YES/NO Prey Dynamics Daphnia density prior to recruitment of young instars Lake > Enclosures LOW > HIGH YES YES Solitary rotifer density prior to recruitment of young instars Lake < Enclosures LOW < HIGH YES NO 129 enclosures did not reach higher densities than the low treatments, contrary to expectation. Finally, by August 1992, most enclosures contained lower densities of prey than the lake and higher densities of Chaoborus. 130 3.4 DISCUSSION Daphnia females hatching out of resting eggs in the spring, also called exephippial females, are the product of sexual reproduction in previous seasons. They have tremendous growth and reproductive abilities, providing this group with intrinsic growth rates 2-4 fold larger, in the case of Daphnia longispina and D. galeata, than the subsequent generations of parthenogenetic females (Arbaciauskas and Gasiunaite 1996). Exephippial Daphnia mature earlier, have clutch size that are 3 to 4 times larger than those of parthenogenetic females. Moreover, the former produce offspring which themselves will grow faster and mature earlier than offspring from parthenogenetic females from later generations. This combination of characteristics results in higher population growth than what can be achieved under equivalent environmental conditions by parthenogenetic females born a few generations later. When Chaoborus predation on exephippial females and their progeny in the spring is high, predator impact could reduce Daphnia densities and could lower the population growth rate. This could result in substantial delays in reaching the Daphnia population maximum, and can also result in a lower maximum density for the season. If exephippial Daphnia rosea also possess these high growth and reproduction characteristics, this could explain the differences in delays of population onset and maxima observed between the different predation treatments within a year. 131 The timing of Chaoborus predation in spring seems to have had at least as large an effect on Daphnia dynamics as the numerical losses imposed by predation. In 1993, lower predator densities in the spring than started experimentally in the previous spring apparently resulted in similar or greater delays in Daphnia population growth, and Daphnia density reduction, than in 1992. Enhanced predator densities in 1992 were experimentally applied after the Daphnia population had had a chance to increase in density in May. By contrast, the predator density levels in spring 1993 had already been in place, via natural recruitment from the previous year, throughout the winter. For example, the predator density in enclosure H was enhanced to over 400 m" in spring 1992, at the end of May. In spring 1993, only about 150 Chaoborus m"3 were present, but this as early as April. Daphnia population density at the time when first instar Chaoborus appeared, was larger in 1992, than in 1993 (1000 Daphnia m" vs. <10 Daphnia m" , respectively). Increased levels of predation in early spring potentially removed Daphnia individuals providing the highest potential for fast population growth which would have changed the Daphnia densities at the onset of population growth. Moreover, the spring predation impact could have been felt later in the season, even after the predators had pupated, because predation might also have changed Daphnia population growth rate and timing of maximal densities. The delays in Daphnia population growth are essential to provide a window of high abundance of small prey species to promote and enhance recruitment of 132 young predators. At low temperatures, predation impact was increased on recruiting young Daphnia, which generated delays in population growth (Neill 1981b). Under colder conditions, Daphnia reproductive rate declines (Orcutt and Porter 1983), and the young Daphnia developmental rate (Hebert 1978; Neill 1981a; Berberovic et al. 1990) and growth rate also decrease (Orcutt and Porter 1983) which keep the young Daphnia in size categories sensitive to invertebrate predation for a longer period of time. Thus, in colder water fourth instar Chaoborus could delay the onset of Daphnia population increase and slow down their population growth by eating a greater proportion of Daphnia recruitment (Neill 1981a). Planktivorous fish, such as Cisco, can also delay Daphnia population increase and the timing of its maximum density because at low temperature the feeding rate of the predator exceeds Daphnia's recruitment rate (Rudstam et al. 1993). In low predation enclosures in Shirley Lake in 1992, abundance of Daphnia populations were delayed temporarily, but finally converged to lake density levels later in the summer. The increase in Daphnia population was sufficiently delayed to release rotifers from competition and improve recruitment of young predators. Similarly, in the Gwendoline Lake experiments, spring predation in a cool year delayed, but did not eliminate, the timing of food limitation for Daphnia (Neill 1981a); zooplankton biomass was reduced and young predator recruitment improved (Neill 1988a). If the predator levels were able to reach this enhanced predation level every spring, a persistent Chaoborus state would occur 133 provided sufficient prey were subsequently available to third and fourth instar predators. Alternatively, in a cold spring year, the same density of predators on a slowly growing Daphnia population could generate a larger predator impact which could also result in negative feedback on the predator dynamics if prey became too scarce. This presents a different mechanism than the one observed in high predation experiments by which Daphnia population would not reach density levels similar to the lake. This mechanism could also lead to low prey availability for third and fourth instars resulted in predator developmental delays. Thus, the persistence of the Chaoborus state relies on a limited loss of Daphnia population resilience. The loss must be sufficient to lower, Daphnia densities at the time when Chaoborus first and second instar recruit, but the loss cannot be so high as to result in food shortage for third and fourth instar growth and development. The negative feedback of high predator densities establishes an upper limit on the Chaoborus densities, i.e. on the upper end of the potential Chaoborus domain. Above this upper limit the system crashes. Over the three year time period recorded for the experiments, all the enclosures lost their predator cohort although enclosures failed at different times. Enclosure dynamics were comparable to lake dynamics early in the experiment but dynamics became more difficult to interpret and less comparable to the lake dynamics in the later part of the experiment as enclosure failure increased. 134 During the first half of the experiment, Daphnia dynamics in low predator treatments and in the lake are similar with a fast increase phase after fourth instar larvae predators have pupated, and with high Daphnia densities reached in the summer. In the latter part of the experiment, only a few enclosures maintained viable Daphnia populations which, in the absence of predator recruitment, reached high densities throughout the summer and fall. In the enclosures with predators still present, delays in development incurred under experimental conditions might interfere with the normal life history of C. americanus , for example, through cannibalism of delayed fourth instar development on recruiting young predators. The sequence in species recruitment failure and the timing of enclosure failures and provided information on the key interactions in the community. Overall, the enclosures provided information about the ability of the predator to overwinter in high density and about the limitations to the predator state in oligotrophic conditions. This information will be useful to model the community to explore the dynamics (size and shape) of the predator state over an increasing nutrient gradient. In summary, because the nutrient levels in my enclosures were still low compared to mesotrophic and eutrophic lakes, the balance between the positive and negative feedback generated by reducing Daphnia population through increased predator densities might only occur for a narrow range of disturbances. 135 Beyond that range, the community bounces out of the Chaoborus state to go to extinction or to return to the Daphnia domain. 3.4.2 Rotifer population dynamics Lower Daphnia densities should release rotifers from competition pressure, and bring about higher rotifer densities (Neill 1984; Gilbert 1988a; 1989). With higher predator densities, and associated lower Daphnia densities, the enclosures were expected to display larger densities of rotifers than the lake. The monthly time scale data revealed no such trend (Appendix E), most likely due to the rapid population recovery that rotifers can exhibit (Edmonson 1965; Stemberger and Gilbert 1985; Walz 1995) through their short generation time. At smaller time scales, I observed the expected lower rotifer densities in spring 1992 in the lake and in the enclosures up to day 182 (end of June) (Figure 3.4). Assuming that the rotifers responded to reduced Daphnia competition in the enclosures as Daphnia did to the enhanced predation pressure (Figure 3.6), I expected the highest rotifer densities and the lowest Daphnia densities in the high predation enclosures (Figure 3.11). However, rotifer densities in medium and high predation enclosures were no higher than those in low predation enclosures, but they were higher than in the lake. The lowest densities of Daphnia generated rotifer densities which were lower than expected from linear projections from lake to low predation enclosures. The rotifer density increase in relation to Daphnia low densities was limited by a factor other than Daphnia 136 competition in the medium and high predation enclosures. A potential explanation for this observation is that fourth instar larvae had become less selective when densities of their preferred prey, such as crustaceans, were low (Pastorok 1980a; b). Chaoborids readily feed on rotifers (Fedorenko 1975b; Moore and Gilbert 1987; 1988b), or any other motile prey small enough to be captured and ingested (Moore et al. 1994). Under high spring predation levels, the rotifer assemblage had only a very short period of time, between fourth instar pupation and first and second instar appearance, to increase to sufficient levels to support the new predator recruits. Unless the medium and high predation enclosures are assumed to have had deficient phytoplankton resources, enhanced predation levels in the spring might play a wider role than solely reducing Daphnia populations. That is direct predation mortality may have reduced the benefits of enhancement in reproduction from competitive release that rotifers experienced. After fourth instar predators pupated, the rotifer densities increased quickly but had only a short period of time for population growth before the appearance of first and second instars increased predation pressure again on the rotifer populations. Such a limitation imposed by predation was underlined by the fact that, between day 182 and 210 (Figure 3.4), rotifer densities in the high predation enclosures continued to decrease below 10 000 individuals m" while they increased substantially in the lake (to 46 000 rotifer individuals m" ) and in the low predation enclosures (up to 38 000 rotifer individuals m" ), even under relatively 137 higher Daphnia densities in the lake and the low Chaoborus treatments. Predation rather than Daphnia competition probably limited rotifer population increases in high spring predator treatments. After predation by Chaoborus was removed, through emergence of fourth instars, followed by young predator recruitment failure, two enclosure communities out of the six which persisted till 1993 or 1994, showed relationships representative of the competition interaction between Daphnia and the solitary rotifers. After their predator cohort failed, enclosures I and H maintained community interactions representative of the Daphnia domain. 3.5 CONCLUSION The density of fourth instar larvae in the spring influenced Daphnia and solitary rotifer population dynamics for the season. Enhanced predation levels resulted in both immediate and long-lasting effects on the Daphnia population dynamics. Immediate effects included delays in Daphnia population onset, which resulted in lower Daphnia densities at the time when young predator instars hatched and recruited. Longer-term effects resulted in lower Daphnia population densities later in summer. The longer delays, and the long-lasting effect under higher predation were likely due to the fact that in early spring, fourth instar predators were feeding on the Daphnia individuals with the highest potential for rapid growth and reproduction, i.e., the Daphnia females which hatch out of resting ephippial eggs and their progeny. 138 Increases in solitary rotifer densities in relation to reduced Daphnia populations were limited to a short period of time, after which young predator recruitment could apparently depress rotifer numbers. In medium and high predation treatments, during the time when Daphnia population growth was delayed, rotifer densities did not respond so strongly as expected to the release from Daphnia competition. Excess predation in these enclosures, potentially by fourth instar predators earlier in the season and certainly by enhanced densities of young predator recruits, reduced rotifer population growth. Such negative feedback brought about by the impact of enhanced predator densities on the prey community could be an indication of an upper limit to the predation state. 139 CHAPTER 4 CHAOBORUS PUPATION IN COLD WATER: IMPLICATIONS FOR LIFE HISTORY, DISTRIBUTION AND POPULATION DYNAMICS. 4.1 INTRODUCTION The phantom midge, Chaoborus (Diptera: Chaoboridae), is a voracious aquatic predatory insect larva that feeds on zooplankton. The influence of Chaoborus larvae predation, on prey population and community dynamics, has been addressed by several studies (Pastorok 1980a; b; Neill 1981b; Moore 1988b; Christoffersen 1990; Havens 1990). In this chapter, I address the influence of the pupal stage, the non-feeding stage in Chaoborus life history, on the prey dynamics, and indirectly on the recruitment of the young predators. In the spring, Chaoborus larvae feed on Daphnia as they hatch from resting eggs, and Daphnia start their population increase. As the larvae pupate, predation pressure is reduced, and the Daphnia population can increase rapidly, thus reducing rotifers, which are prey for the young predators. The longer the period between the onset of pupation and emergence, the lower is the potential for good predator recruitment, and for persistence of the predation state. Unfortunately, little information is available on Chaoborus pupation, and when available, it relates to pupation in warmer conditions (> 12°C) (Luecke 1988; Christoffersen et al. 1993 b). Considering that at least one generation of nearctic 140 multivoltine populations, and all generations of univoltine populations, will have to undergo pupation and metamorphosis in the relatively cold waters of springtime, little information is available to evaluate the impact of water temperature on pupation. In the laboratory, I show that the impact of temperature can be substantial, but is different for the two species of Chaoborus, C. americanus and C. trivittatus. I relate the influence of temperature to our understanding of the life history of these two species, their geographical distribution, and their potential impact on the prey population and their own recruitment. 4.2 MATERIAL AND METHODS 4.2.1 Field collection and laboratory set up Four experiments were set up (Table 4.1). The first two, at 5°C, were devised to study the winter survival abilities of Chaoborus trivittatus and Chaoborus americanus. The next two, at 9°C and 12 °C, were set up to investigate, in more detail, the limitations of C. americanus development in cold water. In all cases, fourth instar larvae were collected at night, from Shirley Lake, a small oligotrophic fishless lake in the Coast Range mountains of British Columbia, Canada. They were collected using a 102 |xm mesh zooplankton net (diameter: 0.4 m) hauled from 6-8m below the surface, at the deepest spot in the lake. Larvae for the 5°C experiment were captured at the end of the season, shortly before the lake froze, on November 17th, 1992. For the 9°C and 12°C experiments, the larvae were collected in springtime, on May 4th, 1995. After capture, the larvae 141 Table 4.1 Laboratory experimental conditions for raising Chaoborus. Temperature (°C) Light Regime (hr. light: hr. dark) Food Regime Chaoborus Species Dates (Start-End) Number of Larvae at start 5 0:24 weekly C. trivittatus Dec92-Mar94 24 5 0:24 weekly C. americanus Dec92-Mar94 25 9 16:8 every 1-2 days C. americanus May95-Sept95 32 12 16:8 every 1-2 days C. americanus May95-Sept95 32 142 were transported to the laboratory in 20 L carboys, with zooplankton collected in the same hauls, and were put in an incubator at 5°C on arrival (between 1-2 hours after collection). The larvae were left to feed for 48 hours, then separated individually into 250 ml plastic containers containing lake water sieved through a 20 |im mesh net. Only larvae with food in their gut were chosen for the experiments. During the experiment, larvae were provided with small prey, either small Daphnia, nauplii, or copepodites, depending on availability. The 5°C experiment started in early December 1992 and ended in March 1994. Larvae were identified to species and put in 250 ml plastic containers with lake water. The containers were put on trays to facilitate future handling and observations, and were all put into an unlighted incubator (0 hr. light : 24 hr. dark) at 5°C, to simulate winter conditions under ice in the lake. The larvae were taken out once per week and the following observations were made: 1) stage of individual: larva, pupa or adult; 2) alive or dead; 3) food trace in gut or empty gut; 4) the number of prey eaten. Prey (eaten or dead) were replenished as needed. Larvae were returned to the cool unlighted regime incubator as soon as possible (15-45 minutes). The 9-12°C experiments took place from early May 1995 to late November 1995. Only larvae with food in their gut were used. The larvae were assigned randomly to the 9 and 12°C incubators. Larvae were kept individually in lake water in 250 ml plastic containers. The containers were grouped in 4L pails to 143 facilitate handling and observation. The incubators were set with a 16 hr. light: 8 hr. dark cycle. Larvae were generally provided with as much food as they could eat. Zooplankton prey included, depending on the season and source of supply, small and medium size Daphnia, nauplii, small and medium size copepodites of Diaptomus kenai and D. leptopus, and adult copepods of D. leptopus, collected from Shirley lake, or when the mountain lake was inaccessible in winter, from ponds on the UBC campus. Larvae were observed and food replenished every 1-2 days. Observations included: presence/absence of food in the gut, number of prey eaten, signs of approaching pupation or metamorphosis, stage (larva, pupa or adult), alive/dead. 4.2.2 Data analysis methods The proportion of larvae which died, pupated and metamorphosed was analyzed with a % test. The life span, or stage duration data, were explored using medians and box plots, and further analyzed using survival analysis (Pyke and Thompson 1986). I used the Kaplan-Meier product limit (program "Pollock" by Dr. C. J. Krebs, Zoology, UBC) to calculate the survival rates over time, for each experiment (Pollock et al. 1989). Time was defined either in days, or in degree-days, depending on the question investigated. I traced the survival curves, and tested the distribution underlining these curves, using a two sample test (Pyke and Thompson 1986) based on Cox's model (S-Plus 1995). 144 4.3 RESULTS Chaoborus americanus pupated at 5, 9, and 12 °C, with substantial emergence only at 12°C (Table 4.2 ). On the other hand. C. trivittatus, at 5°C, showed quite a different pattern. Over half of the larvae that pupated, emerged. A % test showed no significant difference amongst the experiments, for the number of larvae which died (% = 5.33, df=3, p=0.15). On the other hand, the number of pupae which died versus emerged was highly significant (%2 = 63.55, df=3, p«0.0001). Cold water temperature does not affect the switch from larva to pupa, but can have a drastic effect on the switch from pupa to adult emergence. The median number of days individuals spent as pupae was inversely related to the water temperature (Table 4.3). Moreover, variability in duration of pupae in Chaoborus americanus was also inversely related to temperature. The median number of days spent as pupae at 5°C, was the same (28 days or 140 degree-days), for the two species, but over half of C. trivittatus pupae metamorphosed, while all C. americanus pupae died (Table 4.2). C. americanus showed greater variability in pupation duration compared to C. trivittatus (Table 4.3). The Chaoborus americanus larvae held at 5°C, and those held at 9°C, had survival curves with similar shapes (Figure 4.1). One important difference was 145 Table 4.2 Status of Chaoborus larvae in laboratory experiments at different temperatures. The larvae which pupated but did not emerged died in the pupal stage, except for the one C. americanus pupa still alive when the 5°C experiment was terminated. Number of larvae Experimental Species at start which alive which which Temperature Chaoborus died at end pupated emerged (°C) 5 trivittatus 24 4 0 20 11 5 americanus 25 2 1* 23 0 9 americanus 32 3 1 28 2 12 americanus 32 0 0 32 29 * alive as pupa 146 Table 4.3 Duration of pupation at the individual level Temperature SPECIES N= PUPATION (in days) (median time as pupa) Pupation Min. time Pupation Max. time 5 trivittatus 20 28 21 35 5 americanus 23 28 3.5 32.2 9 americanus 28 19 10 23.1 12 americanus 32 12 10 13 147 Figure 4.1 Comparison of survival rates for Chaoborus americanus larvae raised at 5,9, and 12 °C. Time (degree-days) 148 that the die-off at 5°C took place later on the degree-day scale. These two curves are statistically different (Cox's calculation for two sample test: p=0.0001). The shape of the 12°C curve was different than that of the 5°C and 9°C curves described above (Figure 4.1).The curve ended at approx. 1200 degree-days due to metamorphosis, and only a small proportion of the population died before metamorphosis took place. The decline in survival in both the 9°C and the 12°C curves started at 600 degree-days. However, at 800 degree-days, the two groups bifurcated from one another: 12°C pupae emerged, while the 9°C individuals died. By 1100 degree-days, all surviving 12°C pupae emerged while most 9°C individuals died. Chaoborus americanus death rate at 9°C was constant over the 800-1100 degree-day interval. 149 4.4 DISCUSSION Water temperature can influence aquatic invertebrates directly by affecting rates of individual development, and indirectly, by changing the timing and the strength of interactions in relation to their predators, competitors and prey. In laboratory experiments, I determined that water temperature did not have a significant effect on the number of larvae that reached pupation. The most direct influence of temperature on Chaoborus individuals was in determining the number of pupae which could emerge. At the colder temperature, the numbers which could emerge declined. This effect was different for the two species of Chaoborus. In C. trivittatus, 53% of pupae were able to metamorphose at 5°C, while no C. americanus completed ecdysis (Table 4.2). Development in C. americanus seems to be limited by a temperature threshold, that is C. americanus needs a temperature cue to metamorphose and finish its life cycle. By the time individual pupae reached an accumulation of 800 DD, the level at which 12°C individuals started pupating, the proportion of 9°C individual had declined to 70%. By 1100 DD, when 12°C individual finished pupating, 9°C individual survival had declined to below 20%. Moreover, C. americanus pupae held at 5°C, lived as long (28 days) as those of C. trivittatus, but the former died rather than emerge like the latter. Thus, the accumulation of a certain number of degree-days is not sufficient as a threshold to allow metamorphosis in C. americanus: a minimal temperature threshold is needed. Such a temperature threshold, rather than a degree-day threshold, could explain the difference in geographical 150 distribution between C. americanus and C. trivittatus. The latter is found in lakes in higher latitude (Borkent 1981) and higher altitude (Borkent 1981; Lamontagne et al. 1994) than C. americanus. In relation to its more southerly distribution, Lamontagne et al. (1994) showed that C. americanus was not found in lakes with a mid-summer surface temperature below 16°C. Delays in individual development or increased mortality due to cold temperature lead to delayed onset of pupation and emergence, and can affect population and community dynamics. Cold water temperature delays the onset of pupation and emergence by lengthening the time spent in the different larval instars. Temperature induced developmental delays can indirectly affect the interactions of Chaoborus with their prey and predator populations. Luecke (1988) showed that in Lake Lenore, pupae were captured by trout in greater proportion than the larvae although the latter had relatively larger density. In his study, pupae of C. flavicans emerged in two days at 19.2°C, while they required 8 days at 16°C. He modeled the emergence in relation to temperature and predation by trout, and he calculated that trout could remove 23% of the population of Chaoborus pupae day"1 at the peak of the pupation period. The colder was the water the longer was the pupal stage, which increased the probability for pupae of being eaten before emergence. He suggested that this could explain why the summer cohort in Lake Lenore had better recruitment than the spring cohort. In the presence of a pupa predator, Chaoborus suffers greater losses in cold water. By the same principle, but in regards to Chaoborus 151 predation on Daphnia, as cold water lengthens the pupation period, Daphnia population is released from predation by Chaoborus for a longer period of time. As Daphnia population increases, rotifers are outcompeted and their populations decrease [Neill, 1984 #250; see Chapter 3]. Under prolonged pupation and delayed emergence, predator recruitment could be drastically reduced due to a scarcity of rotifers upon which to feed. This effect is compounded in C. americanus because emergence is prevented until temperature reaches 9 °C or more. In the event of a cold spring, the Daphnia population has a longer window of time to increase its density, if it is little affected itself, by cold temperature. In the laboratory, I showed that C. americanus needs a water temperature above 9°C to produce substantial emergence. In the field, this threshold could be lower. In lakes, C. americanus larvae undergo diel vertical migration (Teraguchi and Northcote 1966; Fedorenko and Swift 1972). Pupae have also been observed to migrate vertically (Luecke, 1988; pers. observ.). (Luecke 1988) By migrating to the upper water layers, pupae take advantage of warmer temperatures to develop faster. Moreover, under fluctuating temperature, the favorable range of temperature for development can be expanded, especially in the lower temperature range (Ratte 1985). Even under temperature regimes showing a daily mean temperature below the lower constant temperature limit, some insects are able to complete their development, while they fail to develop under the mean comparable constant temperature (Lin et al. 1954; Messenger and Flitters 1958: in Ratte, 1985). Figure 4.2 schematically describes the relationship 152 Figure 4.2 Schematic representing the effect of developmental rate acceleration due to fluctuating temperature in the low part of the range of developmental temperature of an insect. As temperature increases, so does developmental rate. However, in the lower part of the range, developmental rates under fluctuating temperature are faster than expected by extrapolating to developmental temperature zero (T0) indicated by the dashed line intercepting the temperature axis (based on Messenger and Flitters, 1958: cited in Ratte, 1985). I L T0 Low High Mean temperature 153 between mean temperature and developmental rate. The dashed line represents the relationship expected under a constant temperature regime. The relationship is extrapolated down to the developmental zero temperature, where no growth takes place. The bold line represents the relationship for a fluctuating temperature regime, with the same mean as the constant regime. Over a certain temperature range, both the constant and the fluctuating regime, produce similar developmental rates. In the higher part of the temperature range, the constant temperature regime produces higher developmental rates, than the fluctuating temperature regime. However, in the lower part of the temperature range, the fluctuating temperature regime produces the higher developmental rates (Ratte 1985). If C. americanus can overcome the fixed 9°C threshold, under fluctuating temperature I would expect that more pupae would be able to emerge than the number observed in my fixed temperature experiments in the laboratory. Ratte (1979; 1985: cited in Buns (1991)) showed that for larvae of C. crystallinus in the laboratory, fluctuating temperature allowed faster development. But in field enclosures, larvae raised under both homogeneous (within the epilimnion) and fluctuating temperatures (where larvae were allowed to migrate between the epilimnion and cooler hypolimnion), had nearly equal developmental times. This result might be confounded, because food levels in the enclosures were lower than in the laboratory experiments, and larvae might have been limited by food, rather than by temperature. He did not investigate the effect of fluctuating 154 temperature on pupa developmental rates. However, as food level is irrelevant during pupation, I would expect that pupa developmental rates would be affected by temperature fluctuation, as in Ratte's laboratory experiments. However, different species can develop at different rates under the same temperature regime. Fluctuating temperature experiments in the laboratory or in the field are needed to determine the impact of such a temperature regime on pupal development. Moreover, physiological and biochemical experiments are needed to pinpoint why C. americanus pupae can develop and darken as if to emerge, but cannot complete pupal ecdysis in cold water. Perhaps certain essential chemicals are not synthesized or active below the threshold temperature. 4.5 CONCLUSION Water temperature below 9°C prevents Chaoborus americanus emergence in the laboratory. This partly explains this species' more southerly and lower altitudinal distribution, compared to C. trivittatus. In the field, pupae would encounter fluctuating temperature through diel vertical migration, and could potentially avoid the limitation imposed on their development by the fixed temperature threshold observed in the laboratory. The minimum effect of cold water temperature on Chaoborus pupal development is to lengthen pupation duration, and to delay emergence. This prolongs the period of time available to Daphnia to bloom under low predation pressure. This could have a negative impact on the resilience, and the persistence of the predation state. 155 CHAPTER 5 GENERAL DISCUSSION AND CONCLUSION 5.1 The makings of an extended transient state Shirley Lake does not have two functional domains of attraction of its zooplankton community in relation to predation by Chaoborus. In theory, the Chaoborus domain could still be present, but it is apparently very small. In practice, considering the variability in environmental factors, the community would not stay within such a small domain long enough for this domain to be detectable. The community would barely have time to settle into the Chaoborus domain before being pushed out of it again by a new disturbance. Delays in predator development induced by high predator densities in the spring support the idea that no, or at most, a small Chaoborus domain exists in oligotrophic lakes such as Shirley Lake. Enclosure experiments containing high predator densities indicated that prey populations could be depleted. Thus, the prey community production rate could be overcome by the Chaoborus predation rate. One indication of limited prey production was observed as reductions in both Daphnia and rotifer densities in medium and high predator treatments compared to their densities in the low predation experiments and in the lake. Another indication of limitation was 156 delays in predator development, which suggested that prey production was not sufficient to feed the enhanced predator populations in late summer. Resistance mechanisms, such as flexible predator growth, can usually counterbalance variability and limitation in prey production and prevent the system from switching state. For example, in fish, delays in individual growth such as stunting are observed, especially in older year classes, when food is sufficient for survival but hardly for growth (Scheffer et al. 1995). Chaoborus also show developmental delays under low food conditions (see Chapter 2, section 2.3.3). In the Shirley Lake field enclosures, such starvation-resistance mechanisms were overcome by the strength of the disturbance and recruitment of older instars was lower than necessary to sustain predation pressure. Although young instar recruitment had been improved, large prey were not available in sufficient densities for all the predator larvae that had survived the enlargement of the original recruitment bottleneck of young predators. In other words, the bottleneck moved from first and second instars to the third instar (Figure 5.1). Survival bottlenecks are not an attribute of the predator, Chaoborus, or of the zooplankton community of Shirley Lake exclusively. In general, animals and plants that live in variable or unpredictable environments hedge their bet and produce large quantities of young. In adverse conditions, young recruits encounter a limitation, e.g., food, territory or egg laying site availability. Few will survive to reproduce. In more favorable environments, the bottleneck can be overcome and survival is improved. Organisms with complex life histories, 157 Figure 5.1 Schematic representation of the bottlenecks in Chaoborus recruiment. An increase in nutrient moves the start of the bottleneck without much enlargement (a, b). Only when nutrients are increased sufficiently to increase substantially zooplankton productivity can the bottleneck be enlarged. Chaoborus Instar a) oligotrophic lake (e.g. Shirley Lake, and Gwendoline Lake, Neill, 1988b) b) enclosures with added nutrients c) eutrophic lake (e.g. Triangle Lake, Havens, 1990) 158 including those with ontogenetic shifts, face the possibility of a bottleneck at each stage in their development. Good environmental conditions at one stage does not prevent the potential for a bottleneck at another stage in their life history. In these organisms, the bottleneck can only be overcome if environmental conditions are favorable at all stages of their life history. In my enclosures, the bottleneck in Chaoborus survival took place at the third instar. However, delays in predator development from third to fourth instar did not apparently reduce overwinter survival of fourth instars but could have lengthened the time they spent as fourth instar larvae in the spring and could have desynchronized the timing of pupation and reproduction. Presence of fourth instar larvae over longer periods of time in the spring delays Daphnia population growth and should improve recruitment conditions for young predator larvae. However, as older larvae can cannibalize the young recruits (Fedorenko 1975b) there could be an appreciable reduction of young predator recruitment, despite low Daphnia densities and high rotifer densities. Fourth instar larvae with developmental delays in spring can negate the benefit provided by delayed Daphnia population increase and prevent persistence of the Chaoborus state. In my experimental enclosures, the threshold was initially crossed and the community dynamics moved into the Chaoborus state. This alternative state did not persist over several predator generations: it is thus a transient state. 159 However, the predation state was maintained for more than a season. The signal generated by the experimental predator gradient imposed in 1992 was transmitted overwinter, that is from one season to the next. Enhanced fourth instar predator in spring 1992 resulted, in spring 1993, in enhanced fourth instar predator density which generated important delays in Daphnia population growth. Aquatic ecologist often think that winter resets the dynamics to zero for the start of a new growth season. In my experiments, predator densities were not reset to low level as in the lake after the winter. The predator state persisted for more than one season but less than two predator generations. The predator state is thus an extended transient state. From an ecological perspective, extended transients can provide an opportunity to maintain high diversity in an ecosystem by allowing inferior competitors or rare species to take advantage of the change in conditions. For example, in Shirley Lake enclosures, the transient state allowed longer windows of recruitment for rotifers, Chaoborus, and assumably for phytoplankton. The time duration of an extended transient state depends on the generation time of the longest-lived main player in the system. For example, in Lake Mendota, a strong year class (1977) in Cisco, a long-lived planktivore fish, reduced the population of Daphnia pulicaria, a previously dominant zooplankton prey (Rudstam et al. 1993). The latter was replaced by Daphnia galeata, a smaller species, less sensitive to size-dependent predation by fish. The switch to the smaller and more starvation-prone suspension feeder led to shorter clear-water phases in the 160 spring. Phytoplankton dynamics and Daphnia composition and biomass were affected for 10 years by the success of one year class of a dominant planktivore (Rudstam et al. 1993). The shift between alternative states, be they domains or transient states, occurs in systems where one or more of the functionally "important" species work at the edge of their ability. Working in what could be considered an extreme condition (e.g. limited resource, limited refugia, or physiological stress) for that species, the system will move in a new direction (Grimm and Wissel 1997). The system approaches and crosses a boundary between states. The system dynamic moves from density-dependent to density-vague behavior. For a population with a sufficiently large variance in numbers, recruitment will seem to be density-vague at intermediate levels, and density-dependent at small and large population levels (Carpenter 1988b). 5.2 Dynamical thresholds: the role of nutrient availability, temperature and species composition In many experiments, increased nutrient availability has been required to obtain the shift in community dynamics (Neill 1988a; Moss 1990). Furthermore, whole lake studies (Neill 1988b; Moss 1990; Stenson 1990) show that the mechanisms observed in enclosures also apply at the lake scale. Neill (1988b) hypothesized, based on his enclosure and whole lake experiments in Gwendoline Lake, that oligotrophic lakes do not produce enough prey to sustain higher Chaoborus 161 predator densities. In his whole lake study, fertilization of the lake during one season modified the zooplankton biomass, however it did not result in a lasting qualitative change, and did not increase predator densities in the subsequent season. Stenson (1990) limed a fishless, acidic, oligotrophic lake and observed a switch to increased Chaoborus (C. flavicans and C. obscuripes) densities. Liming in acidified lakes improved abiotic properties of acid water and freed up nutrients. Small cell phytoplankton density, rotifer density, and predator recruitment increased. On the other hand, Bosmina coregoni, a small cladoceran (< 0.5 mm in length), decreased. The switch to Chaoborus state persisted over two years in Gardsjon Lake in Stenson's study. In Shirley and Gwendoline Lakes, Chaoborus were present in low densities and had little impact on cladocerans. Enclosure experiments (Neill and Peacock 1980; Neill 1981b; 1988a) showed that Chaoborus could reduce Daphnia and other crustacean species' population biomass, when the system's nutrient levels were increased. This allowed a greater proportion of the primary production to reach the rotifers, thus yielding an enhanced prey base for young predator recruitment. Predation impact by Chaoborus on prey populations in eutrophic conditions differs from the predator impact in oligotrophic conditions. Havens (1990) conducted exclosure and enclosure experiments in fishless, rotifer-dominated and eutrophic Triangle Lake. Although his enclosures harbored high densities of Chaoborus: (200-800 second and third instars m" in May and June), he concluded, 162 when he compared enclosure and exclosure experiments, that Chaoborus in temperate eutrophic lakes, such as Triangle Lake, do not have a major top-down effect on rotifers. The rotifers' intense reproductive output may greatly exceed losses due to Chaoborus predation. On the other hand, high densities of Chaoborus might impact the crustacean zooplankton to a greater extent because crustaceans have longer generation times than rotifers (Gannon and Stemberger 1978 cited in Havens 1990). In comparison, although second and third instars in Shirley Lake enclosures reached similar levels (200-1000 individuals m"3) as in Triangle Lake enclosures, the former were associated with a reduction in rotifer densities. This is evidence supporting the limitation of rotifer production by low nutrient in Shirley Lake experiments. References about mesotrophic lakes showing multiple domains of attraction have not been found. If these lakes often switched, the data would be very variable, difficult to analyze, and thus might not have been presented in the literature. On the other hand, for similar nutrient levels, mesotrophic lakes are sometimes classified as "oligotrophic" or "eutrophic" based on their species assemblage. Re-analysis, with the idea of thresholds in mind, might provide new insights on the variability in dynamics of mesotrophic lakes. New experiments in mesotrophic lakes might highlight discontinuous system behaviors. Other factors, such as water temperature, can also indirectly reroute primary production to rotifer populations by affecting rates (e.g. growth, predation, 163 survival) differentially between species or groups (e.g. cladocerans versus rotifers, predator versus prey). In a cold year, because Daphnia productivity was reduced more than that of the rotifers, Chaoborus predation impact could be even more substantial (Neill 1988a). Delayed Daphnia dynamics indirectly increase the size of the window for young predator recruitment. Delayed predator development has a more ambiguous effect. Small delays would increase the young predator recruitment window by further delaying Daphnia population. On the other hand, long delays would decrease that window by desynchronizing the predator reproductive period and allowing cannibalism on the young recruits. Change in temperature is not essential to obtain a shift in community structure or dynamic. However, it is an important factor to take into consideration when trying to predict the impact of a disturbance on the community. The same disturbance, applied in a cold or a warm year, can give quite different results (Neill 1988a; Scheffer 1998). Species composition can also influence the state in which the community functions. In the Gardsjon Lake study (Stenson 1990) mentioned previously, minimum and maximum total phosphorus in the lake (3-9 pig l"1 P) includes the range observed in Gwendoline Lake (3-6 |ig l"1 P) and in Shirley Lake (4-7 tig l"1 P). A major difference with Gwendoline and Shirley Lakes is that the dominant cladocerans in Gardsjon Lake, studied by Stenson, were small species, such as Bosmina coregoni and Diaphanosoma brachyurum, rather than large cladocerans such as Daphnia. The small cladoceran species do not monopolize as 164 much of the phytoplankton resource as would larger cladocerans, such as Daphnia (Bogdan and Gilbert 1982; DeMott 1982). In the presence of a smaller and less efficient cladoceran species, rotifers can harvest a greater portion of the primary production to maintain their populations. This means that the presence of the second domain depends not only on the nutrient level that sets the productivity of the system, but also on how this production is divided between the competing prey species. Enclosure experiments using Bosmina in conditions representing Shirley Lake, or using Daphnia in conditions representing Gardsjon Lake would give an interesting comparison between oligotrophic systems. Hence, lake productivity in itself is not a sufficient indicator to predict in which domain or state the lake will be found. The allocation of the primary productivity to the rotifer compartment would be a better indicator of the domain or state in which the community functions. 5.3 Alternative domains and states: importance of the threshold perspective The controversy between Connell and Sousa (1983) and Sutherland (1981; 1990) about the existence of alternative states in natural communities is better understood as a question of perspective. Connell and Sousa did not have a threshold perspective in mind. They considered only systems which responded to the pressure applied to them. Systems which resisted change were not included in their analysis of the presence of alternative states (Sutherland 1990). 165 If a system sometimes remained unchanged and sometimes changed under disturbance, Connell and Sousa only retained as data the system which responded. They had eliminated, by definition, and as a matter of perspective, information pertinent to their discussion. From a threshold perspective, the fact that a system does not respond under certain circumstances and responds under others is of major importance. The work of Connel and Sousa (1983), the criticism of Sutherland (1981; 1990) and the criteria used to choose variables and their range for a study indicate that notions of "stability" (e.g. persistence, resilience, resistance, domain of attraction) are context related. These notions depend on the variables and their range chosen for the study, on the time/spatial scales involved, on the level of description chosen (e.g. population, community), and finally, on the characteristics of the disturbance (Pahl-Wostl 1995; Grimm and Wissel 1997), elements which should be included in a "stability statement" so that comparisons between systems can be made more easily. With a better understanding of what facet(s) of stability is(are) addressed in a study and with more comparisons between systems, thresholds and alternative states might be observed in more systems than those currently. 166 5.4 Hysteresis: one threshold when going up, another when going down Because ecological systems have built-in mechanisms to resist change from one domain to another, it is expected that ecological systems with a threshold crossed when going from domain "A" to domain "B" might have a second threshold when going from domain "B" to domain "A" (Figure 5.2). Hysteresis loops are usually present in models for systems with alternative states (Noy-Meir 1975; Ludwig et al. 1978; Adams and DeAngelis 1987; Scheffer 1990; 1991a; Carpenter and Pace 1997). The presence of a hysteresis indicates that the system possesses resistance, which prevents or delays a change between domains of attraction. It also indicates the range of the variables or parameters at which both the resilience and the resistance of the system are low: where a system's behavior in the face of disturbance will be less predictable. In phase space, this is represented by the area between and including the thresholds forming the hysteresis loop (Figure 5.2). For a disturbance of a certain size, a community outside the hysteresis area remains in the vicinity, or rebounds to its original attractor. For the same size disturbance, a community in the hysteresis area could switch attractor. For example, referring to figure 5.2, apply a two-centimetre disturbance, parallel to the x-axis, to a point sitting on the upper attractor, first under the letter "B", then in the middle of the range where the hysteresis is. In the first case, the point stays on or close to the attractor and its location can easily be predicted. In the second case, the point is taken over a threshold (indicated by xxx). The system will switch attractors. Now imagine that the range where 167 Figure 5.2 Schematic representation of a hysteresis loop and the area of dynamic unpredictability in system with two domains of attraction. The dynamic trajectories of the system are different when the system goes from A to B, and from B to A. The hatched area represents a zone of increased variability and unpredictability in system behavior. In this area, a small disturbance is sufficient to switch the system between the two attractors. 168 hysteresis occurs is not well defined (lots of variability around the location of the thresholds). Redo the second exercise and try to predict in what domain the system will be found. The end result is unpredictable. One can do a similar exercise using the frequency of disturbances rather than size. If the size or the frequency of disturbances, or the true location of the thresholds, are not known or predictable, the resulting domain for a point located away from the hysteresis area will be more easily predicted, than the domain for a point positioned within the hysteresis loop. In the field, separate experiments are required to reveal the two thresholds. For example, I passed the threshold from Daphnia domain to Chaoborus state by adding predators and removing Daphnia. To reveal a hysteresis, I would need to start experiments in Chaoborus state and reduce predators or increase Daphnia densities to see if this threshold is different than the first one. 5.5 Management issues in a threshold perspective From a management perspective, the range of a variable between, and including, the thresholds (see hatched rectangle on variable 1, Figure 5.2) should be avoided entirely to maintain the system within a chosen domain of attraction. In the zone between the thresholds, unpredicted, and unpredictable, events and disturbances will be more likely to bring about large changes in the system than these same events taking place outside the hatched rectangle. From a sampling 169 perspective, the hatched rectangle is the range of the variable that, in the field, will require the use of adaptive sampling designs (Thompson 1992), where the sampling interval changes based on the value of the variable sampled rather than based on the calendar (Ouimet and Legendre 1988). Thresholds associated with extended transient states can become a useful tool in management interventions when managers require a temporary system alteration. The system can be pushed, through a disturbance, into the transient state and it should return on its own towards the attractor of the domain. On the other hand, if a manager requires a "permanent" change in the system and that system possesses an alternative extended transient state rather than an alternative domain, the intervention would look successful at first but would fail over time. When information on the system is scarce, differentiating between a system with alternative domains from one with one alternative domain and an extended transient state is practically impossible. Long-term monitoring of management interventions is essential to collect the information needed to differentiate between the presence of domains and that of extended transient states. 170 CONCLUSION Nonlinearities in prey and predator population dynamics and in predator functional response can lead to thresholds and multiple domains of attraction. Because of such nonlinearities, a predator can regulate prey populations over a certain range of a variable and cannot over a different range. Factors such as size structure and ontogenetic shifts generate pools of individuals with different abilities, thus with different potential growth and population dynamics. Under disturbance, the different pools of individuals (e.g., small versus large, young versus adult) can respond differently, generating nonlinear responses in population dynamics and increasing the possibility for thresholds in system behavior. In variable or unpredictable environments, surplus egg/young production can also generate nonlinearities in dynamics by establishing a high potential for change in population dynamics. This potential can decay over time, resulting in no change in population dynamics, or can be fulfilled, resulting in a switch between system states. In my experiments, enhanced fourth instar Chaoborus densities in the spring expanded the recruitment window for young predator instars. A threshold between the Daphnia domain and the Chaoborus state was crossed. In the literature, the number of systems, aquatic and terrestrial, which are shown to have thresholds in community behavior is increasing, and there are now a few examples of systems with at least two domains of attraction. 171 Considering the potential for nonlinearities in population dynamics and species interactions, and considering the complexity of natural communities, the current paucity of systems representative of multiple domains might be more a reflection of the perspective and tools used to focus research efforts, than a true representation of the functioning of ecological systems. A shift in perspective from 'stability' to variability, from equilibria to thresholds, will influence the results we can observe, the way we interpret them, and more importantly, will influence the questions we can answer (Breckling 1992; Pahl-Wostl 1995). Furthermore, common statistical tools often mask the presence of thresholds. They are well suited for data analysis within states, but can often give erroneous interpretations when the data span more than one domain or state, therefore underestimating the number of ecological systems with thresholds in their dynamics. Once a discontinuity in dynamics has been observed, a distinction between domain of attraction, transient state or extended transient state must be made. This categorization requires detailed information, as well as experimentation and modeling, to determine the extent of the states and their persistence. The distinction between types of states is especially important when designing intervention for ecological management. Determining in which systems, and under what conditions thresholds will be reached, or avoided, is of increasing importance as we steadily increase pressure (rate and size of disturbances) on managed and natural systems. 172 A Enclosure construction design - Enclosures were made of white, woven polyurethane plastic - Float was made of extruded plastic foam and inserted in yellow coated (UV protection) woven polyurethane plastic - Wood frame was made of 2X4 rough cedar 173 APPENDIX B Method for enclosure fill up with pumps Enclosure were filled using 3-inch diameter diaphragm pumps. Five pumps, operated by five volunteers, were used to fill enclosures two at a time, over the same period of time. The pairs were as follows: B-C, D-E, A-F, G-H, I-J. The pump output hose was set in the first enclosure of a pair for about half hour. The hose was then moved to the second enclosure for about one hour. The hose was moved back for another half hour to the first enclosure. In this way, migrating or new blooms of zooplankton could be divided up between enclosures more equitably than by filling one enclosure fully at a time. Water was pumped in until the enclosure was well rounded on the sides. 174 APPENDIX C Comments on predator time series 1992-1994. I present the details of the three year time series for each instar in Appendices C-l to C-4. Time series patterns are based on a single replicate measurement series per station from May 1992 to October 1994. I noted several discrepancies between the life histories of C. americanus and C. trivittatus which can affect the interpretation of the data. C. trivittatus in Shirley Lake seems to have two reproductive periods which means that second instars are present most of the time in the lake. On the other hand, C. trivittatus did not recruit well in the enclosures. The data I present under the expression "total Chaoborus", in the Appendices and in Chapter 2, represent (1) in the lake, the total of both species for each date over the three year study, (2) in the enclosures, the total of both species till July 1992 and only C. americanus from July 1992 thereafter. Several features of Shirley Lake time series originated from the peculiarities in the life history of C. trivittatus. not previously observed in neighboring lakes These included extended period of first instar recruitment from May to November rather than in June-July only (Appendix C-l, panel 1), presence of second instars throughout the year, including a second peak in the fall (Appendix C-2, panel 1). Moreover, I observed in the lake and in some enclosures that C. trivittatus and C. americanus were overwintering in second and third instars (Appendices C-2 and C-3), and that larval development took place over the winter. Finally, first and second instars, when present in the fall (Oct., Nov.), were observed to migrate vertically. They were found in greater density in the deeper samples and in lower density or absent from the 0-3 m 175 samples (Appendices Cl- and C2, panel 1). Usually, young instars are neither found in the fall nor found to migrate. Both the lake and the enclosures experienced not one but two main seasonal declines in fourth instar densities. (Appendix C-4). The first decline, in June and July (Appendix C-4, panel 1)), is representative of the disappearance of fourth instars larvae from the water column as they pupate and emerge to reproduce. The decline in the lake is less pronounced because a good proportion of predators are Chaoborus trivittatus, a species which did not recruit in the enclosures and which, in Shirley Lake, pupate, emerge and reproduce at different times than C. americanus (Figure 2.2). Thus, in the lake, fourth instar larvae are always present, although at low levels (generally < 50 larvae m" ). The second decline takes place in the fall (October-November) and represents not a true decrease in fourth instar larvae densities, but a virtual one. Because the fall and subsequent spring densities are generally similar and because reproduction does not take place in the winter, I attribute the winter decline to a change in "behavior" on the part of the predators. In the lake, where two different depths were sampled (Appendix C-4, panel 1)), I found a greater density of the fourth instar larvae in the deep samples in late fall than during the summer. Assuming the same relationship for the enclosures, I can conclude that the larvae were present, but below the depth sampled. Fourth instar larvae are known to move deeper in late fall and winter (Fedorenko, 1972). Moreover, in cold temperatures, the animals have lower respiration rates (Swift, 1976), and higher food assimilation (Giguere, 1980, 1981). With lower food intake requirement to fulfill in cold conditions, a greater proportion of the population stays in deep water instead of undergoing vertical migration each day. 176 APPENDIX C-l Monthly sample time series for total Chaoborus first instar larva density in Shirley Lake and in experimental enclosures from 1992 to 1994 Time is indicated in months and in Julian days over three years (1) Lake; (2) Low (3) Medium (4) High predator treatments 1992 fr e„ Sj> £ tj fc 2 2, Z< £ o 2 J I I I I L 1993 * fr i- w> "S, H- > I I I I I I I L 1994 I I I 1_J I I I L J I L J I I I I I L 1 1 1 1 r 00000 ^ O CO rH 1—1 1—1 CN CS CS CO CO J I I I I I I L 10 3-i ft |:\ 10 2\ t *** 10 JJ i \ 1 * • io°: I". \ —1—r~i—rn—1—1—r -* fr 60 n, -w > J I I I I I I L I I * fr 1 in 1 r _ SP a- - fc ST e - SP cL — fc illZ$6 I J ^ I I I I L 1 1 1 1 1 1 1 1 fn—1 1 1—1—r 00 00000 ID 00 rH ^< O CO rH r-4 CN CN CN CO CO J I I I I I L 1 1 1—1 1 1—1—r oooooo 00 rH tN O CO rH CN CN CN CO CO J I I I I I L 1—r 1—r 0 o p O J I I I I I I L 1 1 1 1 1 r fr a - SP OH - fc 2 £ £ < cn O Z o. cy t-4 j 3 STUD < S £ Z< % OZ l—rn—1—r ^ r ^ fr r-o- <2 c 177 APPENDIX C-2 Monthly sample time series for total Chaoborus second instar larva density in Shirley Lake and in experimental enclosures from 1992 to 1994 Time is indicated in months and in Julian days over three years (1) Lake; (2) Low (3) Medium (4) High predator treatments 1992 1993 1994 2 £ £ < « O Z j-1 I l__l I I L 0*5 5 'S s « " o a, ,2 a - a' « « o 60 "K, H-. -2 3 ST V> J I I I I I I L J I I I I I I L 10 ^ 10 2-10\ 10°. i i—i i—r i i T—t T—I i—I CN CN CN CO CO _l I I I I I I L_ i—r o — .A f—r o o o o o o o in oo rn  ro rH rH CN CN CN CO CO J I I I I I L o o o o o T-I O CO CN CN (N CO CO J I I I I I I L 10J-J 10' 101 10° "\ I .i "i i ' i—i—r <= 60 Q, > I—TTT—I—T *-3 § i ji' i— » Q, * > "73 3 a> « o 5, < M.OZ i i r T3 3 aj i O r% c3 _ 1V _ i F^/G V- c : ? \ \A A1 / ' .'i i •, / 'A' '* / II i i i i i o o o o o o o ID 00 rH O CO rH rH CN CN CN CO CO 1 l I 1 ' 1 1 11 1 ' 1 oo oooooo (N 1/1 OOrH^txOCO rHrH t-» CN CN CN CO CO 1 l l l l l l l l 00 oooooo CN m COr-lTft>.OCO T—IT—I T-H CN CN CN CO CO 1 i i i i i i i 10 •'-J 10: 101. 10°. 103-J 10 2 101 10° L-3m L-6m \ _ j—i i u i—i r 2 41, < Cfl O Z T^7 *3 TTT I I T^TT I I >-• m a M n, *. > rr i r , ^. . M Q, *» > •72 3 of H O £ < w 0 Z T 178 APPENDIX C-3 Monthly sample time series for total Chaoborus third instar larva density in Shirley Lake and in experimental enclosures from 1992 to 1994 Time is indicated in months and in Julian days over three years (1) Lake; (2) Low (3) Medium (4) High predator treatments 1992 1993 1994 fc fr d 00 a, ** > I I I I I I I L i i i i i r o o o o o 10 3 J I I I I I L J I I I I I I I L J I I I I I I L 10 2 J 101 J F , I IJI I r~TZ1 TTT ^ LI—i—i i~ i—r—T I OH « S •= 3 j u O J I I I L_l I L fc " fr' C ' ' W OH' ' > < £ £ < $ O Z 1 1 1 1 1 1 1 1 I 1 IT 1 11 1 r L 1—1—1—1—1—m oooooo 00000 J I I I I I I L J I I I I I I L 102J 101 J 10' L-3m •£"1 I I hrlV I L I — op rv > 6 ~ * II VfrV '.o'l'I1/ 'fc'fr'd' 1 WD'-^L '> 1 OH « § ~ 3 OH 179 APPENDIX C-4 Monthly sample time series for total Chaoborus fourth instar larva density in Shirley Lake and in experimental enclosures from 1992 to 1994 Time is indicated in months and in Julian days over three years (1) Lake; (2) Low (3) Medium (4) High predator treatments 1992 1993 1994 10" -j V. 101 10° T T—I—TJT—I—r M a. * >* • S 4) w O « « - =? & - -10 3 10 2 101 10° • 103 102 10 10°. J L J L T—i—i—i—i—i—r J I I I I I L 6 m —i—LJ~ i—r~T" 2 £ £•< cn O Z OH « e 60 H- > 1 I I I T—i—i—r ir cs ro ro J I I I I I I L I I " I T—I 1 ' I "l £ & «- 60 "QV H- > J I I I I I I L OH « < 2 T—I—I—I—I—III oooooo OD H ^ K O P) r-H CN CN CN CO CO J I I I I I OH.2 S 60 ^ > m—i—i—i—rn—r OO oooooo N IT) QOrH-^C^OCO i-Hr-t rH CN CN CN CO CO _l I I I I I I L Medium Predation i i i i i i i r J I I I I I I L Low Predation i r i i i i r OO OOOOOO CN in ooi-H-^t^oco rH rH rH CN CN CN CO CO J L J I I I L Lake TTI—T 3 3 I II—|—T 60 fx ^ > 3 57 t3 o (4) (3) (2) (1) 180 APPENDIX D Monthly sample time series for total Daphnia density in Shirley lake and in experimental predator addition enclosures from 1992 to 1994. Time is indicated in months and in Julian days over three years - inverted triangles represent appearance of predator first instars in monthly samples - horizontal guidelines at 1000 Daphnia m3 1992 1993 1994 MJJASON AMJJASON AMJJASON 181 APPENDIX E Monthly sample time series for total solitary rotifer density in Shirley lake and in experimental predator addition enclosures from 1992 to 1994. Time is indicated in months and in Julian days over three years - horizontal guidelines at 10 000 solitary rotifers m"3 1992 1993 1994 Tj CO C cy TJ •3 K -St a. cs Q oooo CN 00 O rH rH CN CO O in 00 Ov CT* io6. io5. IO4. io3. IO2. io1. I I I I I I I I I I I I I I I I I I I I < I 1 ' 1 1 I I I I I I I I M J J AS ON O O O o o CN 00 Tjt O VO rH rH CN CO CO 10 6-10 5-10 4. 10 3-10 2. 101 ' ' I I ' ' I ' L-3m H -V-» i 'rJ-I I I I I I I I I I I AM J J AS ON o o o o o 00 O VO CN in vo vo K I I I I I I I I I I I I I I I I I I I I AM J J AS ON o o o T}< O vO 00 Ov CJv * I I I I I I I I I I I I I I I I I f I I I I I I I I I I I I I I I I I I I I I I MJJASON AMJJASON AMJJASON E TJ S CS E a TJ s It o -cu 182 BIBLIOGRAPHY Adams, S. M. and D. L. DeAngelis (1987). Indirect effects of early bass-shad interaction on predator population structure and food web dynamics. Predation: direct and indirect effects on aquatic communities. C. W. Kerfoot and A. Sih. Hanover, University Press of New England Pp. 103. Addicott, J. F., J. J. M. Aho, M. F. Antolin, D. K. Padilla, J. S. Richardson and D. A. Soluk (1987). Ecological neighborhoods: scaling environmental patterns. Oikos 49: 340-346. Arbaciauskas, K. and Z. R. Gasiunaite (1996). Growth and fecundity of Daphnia after diapause and their impact on the development of a population. Hydrobiologia 320: 209-222. Arndt, H. (1993). Rotifers as predators on components of the microbial web (bacteria, heterotrophic flagellates, ciliates) - a review. Hydrobiologia 255/256: 231-246. Arndt, H., C. Schroder and W. Schese (1990). Rotifers of the genus Synchaeta — an important component of the zooplankton in the coastal waters of the Southern Baltic. Limnologica 21: 233-235. Berberovic, R., K. Bikar and W. Geller (1990). Seasonal variability of the embryonic development time of three planktonic crustaceans: dependence on temperature, adult size and egg weight. Hydrobiologia 203: 127-136. Blindow, I., G. Andersson, A. Hargeby and S. J. Johansson (1993). Long-term pattern of alternative stable states in two shallow eutrophic lakes. Freshwater Biology 30: 159-167. Bogdan, K. G. and J. J. Gilbert (1982). Seasonal patterns of feeding by natural populations of Keratella, Polyarthra, and Bosmina: clearance rates, selectivities, and contributions to community grazing. Limnology and Oceanography 27: 918-934. Borkent, A. (1979). Systematics and bionomics of the species of the subgenus Schadonophasma Dyar and Shannon (Chaoborus, Chaoboridae, Diptera). Quaestiones Entomologicae 15: 122-255. Borkent, A. (1981). The distribution and habitat preferences of the Chaoboridae (Culicomorpha: Diptera) of the Holarctic region. Canadian Journal of Zoology 59: 122-133. 183 Borsheim, K. Y. and S. Andersen (1987). Grazing and food selection by crustacean zooplankton compared to production of bacteria and phytoplankton in a shallow Norvegian mountain lake. Journal of Plankton Research 9: 367-379. Bradshaw, W. E. (1969). Major environmental factors inducing the termination of larval diapause in Chaoborus americanus Johannsen (Diptera: culicidae). Biological Bulletin 136: 2-8. Bradshaw, W. E. (1970). Interaction of food and photoperiod in the termination of larval diapause in Chaoborus americanus (Diptera: culicidae). Biological Bulletin 139: 476-484. Breckling, B. (1992). Uniqueness of ecosystems versus generalizability and predictability in ecology. Ecological Modelling 63: 13-27. Brooks, J. L. and S. I. Dodson (1965). Predation, body size, and the composition of plankton. Science 150: 28-35. Buns, M. and H. T. Ratte (1991). The combined effects of temperature and food consumption on body weight, egg production and developmental time in Chaoborus crystallinus De Geer (Diptera: Chaoboridae). Some new evidence for the adaptive value of vertical migration. Oecologia 88: 470-476. Burns, C. W. and J. J. Gilbert (1986a). Direct observations of the mechanisms of interference between Daphnia and Keratella cochlearis. Limnology and Oceanography 31: 859-866. Carpenter, S. R., Ed. (1988a). Complex interactions in lake communities. New York, Springer-Verlag Pages. Carpenter, S. R. (1988b). Transmission of variance through lake food webs. Complex interactions in lake communities. New York, Springer-Verlag Pp. 119-135. Carpenter, S. R. and K. L. Cottingham (1997). Resilience and restoration of lakes. Conservation ecology [online] http://www.consecl.org/voll/issl/artl 1. Carpenter, S. R. and J. F. Kitchell (1993). The trophic cascade in lakes. New York, Cambridge University Press Pages. Carpenter, S. R. and M. L. Pace (1997). Dystrophy and eutrophy in lake ecosystems: implications of fluctuating inputs. Oikos 78: 3-14. Christoffersen, K. (1990). Evaluation of Chaoborus predation on natural populations of herbivorous zooplankton in a eutrophic lake. Hydrobiologia 200/201: 459-466. 184 Christoffersen, K., B. Riemann, A. Klysner and M. Sondergaard (1993). Potential role of fish predation and natural populations of zooplankton in structuring a plankton community in eutrophic lake water. Limnology and Oceanography 38: 561-573. Connell, J. H. and W. P. Sousa (1983). On the evidence needed to judge ecological stability or persistence. The American Naturalist 121: 789-824. DeMott, W. R. (1982). Feeding selectivities and relative ingestion rates of Daphnia and Bosmina. Limnology and Oceanogoraphy 27: 518-527. Dublin, H. T. (1995). Vegetation dynamics in the Serengeti-Mara ecosystem: the role of elephants, fire and other factors. Dynamics, management and conservation of an ecosystem. A. R. E. Sinclair and P. Arcese. Chicago, The University of Chicago Press Pp. 71-90. Dublin, H. T., A. R. E. Sinclair and J. McGlade (1990). Elephants and fire as causes of multiple stable states in the Serengeti-Mara woodlands. Journal of Animal Ecology 59:1147-1164. Edmonson, W. T., Ed. (1959). Fresh-water biology. New York, John Wiley and Sons Pages. Edmonson, W. T. (1965). Reproductive rate of planktonic rotifers as related to food and temperature in nature. Ecological Monographs 35: 61-111. Edmonson, W. T. (1971b). Methods for processing samples and developing data. A manual on methods for assessment of secondary productivity in fresh waters. W. T. Edmonson and G. G. Winberg. Oxford, Blackwell Scientific Publications. 17 Pp. 127-169. Edmonson, W. T. and J. T. Lehman (1981). The effect of changes in the nutrient income on the condition of Lake Washington. Limnology and Oceanography 26: 1-29. Edmonson, W. T. and A. H. Litt (1982). Daphnia in Lake Washington. Limnology and Oceanography 27: 272-293. Egerton, F. N. (1973). Changing concepts in the balance of nature. Quarterly Review of Biology 48: 322-350. Fedorenko, A. Y. (1975a). Feeding characteristics and predation impact of Chaoborus (Diptera, Chaoboridae) larvae in a small lake. Limnology and Oceanography 20: 250-258. 185 Fedorenko, A. Y. (1975b). Instar and species-specific diets in two species of Chaoborus. Limnology and Oceanography 20: 238-249. Fedorenko, A. Y. and M. C. Swift (1972). Comparative biology of Chaoborus americanus and Chaoborus trivittatus in Eunice Lake, British Columbia. Limnology and Oceanography 17: 721-730. Fretwell, S. D. (1987). Food chain dynamics: the central theory of ecology? Oikos 50: 291-301. Gannon, J. E. and R. S. Stemberger (1978). Zooplankton (especially crustaceans and rotifers) as indicators of water quality. Transactions of the American Microscopy Society 97:16-35. Giguere, L., A. (1980a). Metabolic expenditure in Chaoborus larvae. Limnology and Oceanography 25: 922-928. Giguere, L. A. (1981). Food assimilation efficiency as a function of temperature and meal size in larvae of Chaoborus trivittatus (Diptera: Chaoboridae). Journal of Animal Ecology 50: 103-109. Gilbert, J. J. (1985). Competition between rotifers and Daphnia. Ecology 66: 1943-1950. Gilbert, J. J. (1988a). Supression of rotifer populations by Daphnia : a review of the evidence, the mechanism, and the effect on zooplankton community structure. Limnology and Oceanography 33: 1286-1303. Gilbert, J. J. (1989). The effect of Daphnia interference on a natural rotifer and ciliate community. Short-term bottle experiments. Limnology and Oceanography 34: 606-617. Gilbert, J. J. and J. D. Jack (1993). Rotifers as predators on small ciliates. Hydrobiologia 255/256: 247-253. Gilbert, J. J. and R. Stemberger, S. (1985). Control of Keratella populations by interference competition from Daphnia. Limnology and Oceanography 30: 180-188. Gilpin, M. E. and T. J. Case (1976). Multiple domains of attraction in competition communities. Nature 261: 40-42. Gliwicz, Z. M. (1985). Predation or food limitation: an ultimate reason for extinction of planktonic cladoceran species. Archiv fur Hydrobiologie Beiheft Egebnisse der Limnologie 21: 419-430. 186 Goldspink, C. R. and D. B. Scott (1971). Vertical migration of Chaoborus flavicans in a Scottish loch. Freshwater Biology 1: 411-421. Grimm, V. and C. Wissel (1997). Babel, or the ecological stability discussions: an inventory and analysis of terminology and a guide for avoiding confusion. Oecologia 109: 323-334. Gude, H. (1988). Direct and indirect influences of crustacean zooplankton on bacterioplankton of Lake Constance. Hydrobiologia 159: 63-73. Harrison, G. W. (1986). Multiple stable equilibria in a predator-prey system. Bulletin of Mathematical Biology 48: 137-148. Havens, K. E. (1990). Chaoborus predation and zooplankton community structure in a rotifer-dominated lake. Hydrobiologia 198: 215-226. Hebert, P. D. N. (1978). The population biology of Daphnia (Crustacea, Daphnidae). Biological Review 53: 387-426. Hik, D. S., R. L. Jefferies and A. R. E. Sinclair (1992). Foraging geese, isostatic uplift and asymmetry in the development of salt-marsh plant communities. Journal of Ecology 80: 395-406. Hillbricht-Ilkowska, A., Z. Kajak, J. Ejsmont-Karabin, A. Karabin and J. Rybak (1975). Ecosystem of Milolajskie Lake: the utilization of the consumers' production by invertebrate predators in pelagic and prodoundal zones. Polish Archives of Hydrobiology 22: 53-64. Hoffmann, K. H., Ed. (1985). Environmental physiology and biochemistry of insects. Berlin, Springer-Verlag Pages. Holling, C. S. (1973). Resilience and stability of ecological systems. Annual Review of Ecology and Systematics 4: 1-24. Holling, C. S. (1992). Cross-scale morphology, geometry, and dynamics of ecosystems. Ecological Monographs 62: 447-502. Holt, R. (1977). Predation, apparent competition, and the structure of prey communities. Theoretical Population Biology 12: 197-229. Hunter, M. D. (1992b). The relative contributions of top-down and bottom-up forces in population and community ecology. Ecology 73: 724-732. Jack, J. J. and J. J. Gilbert (1993). Susceptibilities of different-sized ciliates to direct suppression by small and large cladocerans. Freshwater Biology 29: 19-29. 187 Jeppesen, E., J. P. Jensen, P. Kristensen, M. Sondergaard, E. Moretensen, O. Sortkjaer and K. Olrik (1990). Fish manipulation as a lake restoration tool in shallow, eutrophic, temperate lakes. 2: Threshold levels, long-term stability and conclusions. Hydrobiologia 200/201: 219-227. Jorgenson, C. B. (1966). Biology of suspension feeding. Oxford, Pergamon Press Pages. Jiirgens, K. (1994a). Impact of Daphnia on planktonic microbial food webs - a review. Marine Microbial Food Webs 8: 295-324. Jiirgens, K., H. Arndt and K. O. Rothhaupt (1994b). Zooplankton mediated changes of microbial food web structure. Microbial Ecology 27: 27-42. Kajak, Z. and B. Ranke-Rybicka (1970). Feeding and predation efficiency of Chaoborus flavicans Meigen (Diptera: Culicidae) larvae in eutrophic and dystrophic lakes. Polish Archives of Hydrobiology 17: 225-232. Kajak, Z. and J. Rybak (1979). The feeding of Chaoborus flavicans Meigen (Diptera: Chaoboridae) and its predation on lake zooplankton. Internationale Revue der Gesamten Hydrobiologie 64: 361-378. Kingsland, S. E. (1985). Modeling nature: Episodes in the history of population ecology. Chicago, The University of Chicago Press Pages. Lamontagne, S., D. B. Donald and D. W. Schindler (1994). The distribution of four Chaoborus species (Diptera: Chaoboridae) along an elevational gradient in Canadian Rocky Mountain lakes. Canadian Journal of Zoology 72: 1531-1537. Lampert, W., W. Fleckner, H. Rai and B. Taylor (1986). Phytoplankton control by grazing zooplankton: a study on the spring clear-water phase. Limnology and Oceanography 31: 4718-490. LaRow, E. J. (1968). A persistent diurnal rhythm in Chaoborus larvae. I. The nature of the rhythmicity. Limnology and Oceanography 13: 250-256. Leibold, M. A. (1989). Resource edibility and the effects of predators and productivity on the outcome of trophic interactions. The American Naturalist 134: 922-949. Lewontin, R. C. (1969). The meaning of stability. Diversity and stability in ecological systems. New York, Brookhaven National Laboratory Pp. 13-24. Lin, S., A. C. Hodson and A. G. Richards (1954). An analysis of threshold temperature for development of Oncopeltus and Tribolium eggs. Physiological Zoology 27: 287-311. 188 Ludwig, D., D. D. Jones and C. S. Holling (1978). Qualitative analysis of insect outbreak systems: the spruce budworm and forests. Journal of Animal Ecology 47: 315-332. Ludwig, D., B. Walker and C. S. Holling (1997). Sustainability, stability and resilience. Conservation ecology [online] http://www.consecl.org/voll/issl/artl 1. Luecke, C. (1988). Predation by cutthroat trout on Chaoborus flavicans as revealed by a population budget. Limnology and Oceanography 33: 1254-1265. McQueen, D. J., M. Johannes, P. J., S. T. and L. D. (1989). Bottom-up and top-down impacts on freshwater pelagic community structure. Ecological Monographs 59: 289-309. Messenger, P. S. and N. E. Flitters (1958). Effect of constant temperature environments on the egg stage of three species of Hawaiian fruit flies. Annals of the Entomological Society of America 51: 109-119. Mittelbach, G. G., A. M. Turner, D. J. Hall and J. E. Rettig (1995). Perturbation and resilience: a long-term, whole-lake study of predator extinction and reintroduction. Ecology 76: 2347-2360. Moore, M., V. and J. J. Gilbert (1987). Age-specific Chaoborus predation on rotifer prey. Freshwater Biology 17: 223-236. Moore, M., N. D. Yan and T. Pawson (1994). Omnivory of the larval phantom midge (Chaoborus sp.) and its potential significance for freshwater planktonic food web. Canadian Journal of Zoology 72: 2055-2065. Moore, M. V. (1988a). Density-dependent predation of early instar Chaoborus feeding on multispecies prey assemblages. Limnology and Oceanography 33: 256-268. Moore, M. V. (1988b). Differential use of food resources by the instars of Chaoborus punctipennis. Freshwater Biology 19: 249-268. Morley, F. H. W. (1966a). Stability and productivity of pastures. Proceedings of the New Zealand Society of Animal Production 26: 8-21. Morley, F. H. W. (1966b). The biology of grazing management. Proceedings of the Australian Society of Animal Production 6: 127-136. 189 Moss, B. (1990). Engineering and biological approaches to the restoration from eutrophication of shallow lakes in which aquatic plant communities are important components. Hydrobiologia 200/201: 367-377. Neill, W. E. (1981a). Developmental responses of juvenile Daphnia rosea to experimental alteration of temperature and natural seston concentration. Canadian Journal of Fisheries and Aquatic Sciences 38: 1357-1362. Neill, W. E. (1981b). Impact of Chaoborus predation upon the structure and dynamics of a crustacean zooplankton community. Oecologia 48: 164-177. Neill, W. E. (1984). Regulation of rotifer densities by crustacean zooplankton in an oligotrophic montane lake in British Columbia. Oecologia 61: 175-181. Neill, W. E. (1985). The effects of herbivore competition upon the dynamics of Chaoborus predation. Archiv fur Hydrobiologie Beiheft Egebnisse der Limnologie 21: 483-491. Neill, W. E. (1988a). Community responses to experimental nutrient perturbations in oligotrophic lakes: the importance of bottlenecks in size-structured populations. Size-structured populations. B. Ebenman and L. Persson. New York, Springer-Verlag Pp. 236-255. Neill, W. E. (1988b). Complex interactions in oligotrophic lake food webs: the productivity gradient of temperate lake ecosystems. Complex interactions in lake communities. S. R. Carpenter. New York, Springer-Verlag Pp. 31-44. Neill, W. E. and A. Peacock (1980). Breaking the bottleneck: interactions of invertebrate predators and nutrients in oligotrophic lakes. Ecology and evolution of zooplankton communities. W. C. Kerfoot. Hanover, University Press of New England. 3: Special Symposium Pp. 715-724. Noy-Meir, I. (1975). Stability of grazing systems: an application of predator-prey graphs. Journal of Ecology 63: 459-481. Olson, M. H., G. G. Mittelbach and C. W. Osenberg (1995). Competition between predator and prey: resource-based mechanisms and implications for stage-structured dynamics. Ecology 76: 1758-1771. Orcutt, J. D. and Porter (1983). Diel vertical migration by zooplankton: constant and fluctuating temperature effect on life history parameters of Daphnia. Limnology and Oceanography 28. Ouimet, C. and P. Legendre (1988). Practical aspects of modelling ecological phenomena using cusp catastrophe. Ecological Modelling 42: 265-287. 190 Pahl-Wostl, C. (1995). The dynamic nature of ecosystems. Chaos and order entwined. New York, Wiley Pages. Parma, S. (1971). Chaoborus flavicans (Meigen) (Diptera: Chaoboridae): an autoecological study, University of Groningen. Pastorok, R. A. (1980a). Selection of prey by Chaoborus larvae: a review and new evidence for behavioral flexibility. Evolution and ecology of zooplankton communities. W. C. Kerfoot. Hanover, University Press of New England Pp. 538-554. Pastorok, R. A. (1980b). The effects of predator hunger and food abundance on prey selection by Chaoborus larvae. Limnology and Oceanography 25: 910-921. Pastorok, R. A. (1981). Prey vulnerability and size selection by Chaoborus larvae. Ecology 92:1211-1324. Persson, L. (1987). Effects of habitat and season on competitive interactions between roach (Rutilus rutilus) and perch (Perca fluviatilis). Oecologia 73: 170-177. Persson, L. (1988). Asymmetries in competitive and predatory interactions in fish populations. Size-structured populations: ecology and evolution. B. Ebenman and L. Persson. New York, Springer-Verlag Pp. 203-218. Persson, L. and L. A. Greenberg (1990). Juvenile competitive bottlenecks: the perch (Perca fluviatilis) - roach interaction. Ecology 71: 44-56. Peterman, R. M., W. C. Clark and C. S. Holling (1979). The dynamics of resilience: shifting stability domains in fish and insect systems. Population dynamics. R. M. Anderson, B. D. Turner and L. R. Taylor. Oxford, Blackwell Scientific Publications Pp. 321-341. Peterson, B. J., J. E. Hobbie and J. F. Haney (1987). Daphnia grazing on natural bacteria. Limnology and Oceanography 23: 1039-1044. Polis, G. A. (1994). Food webs, trophic cascades and community structure. Australian Journal of Ecology 19: 121-136. Polis, G. A. and D. R. Strong (1996). Food web complexity and community dynamics. The American Naturalist 147: 813-846. Pollock, K. H., S. R. Winterstein, C M. Bunck and P. D. Curtis (1989). Survival analysis in telemetry studies: the staggered entry design. Journal of Wildlife Management 53: 7-15. 191 Power, M. E. (1992). Top-down and bottom-up forces in food webs: do plants have primacy? Ecology 73: 733-746. Pyke, D. A. and J. N. Thompson (1986). Statistical analysis of survival and removal rate experiments. Ecology 67: 240-245. Ratte, H. T. (1979). Tagesperiodische Vertikalwanderung in thermisch geschichteten Gewassern: Einflufi von Temperatur- und Photoperiodezyklen auf Chaoborus crystallinus De Geer (Diptera: Chaoboridae). 57: 1-37. Ratte, H. T. (1985). Temperature and insect development. Environmental physiology and biochemistry of insects. K. H. Hoffman. Berlin Heidelberg, Springer Verlag Pp. 33-65. Riessen, H. P., J. P. Sommerville, C. Chiapparri and D. Gustafson (1988). Chaoborus predation, prey vulnerability, and their effect in zooplankton communities. Canadian Journal of Fisheries and Aquatic Sciences 45: 1912-1920. Rudstam, L., R. C. Lathrop and S. R. Carpenter (1993). The rise and fall of a dominant planktivore: direct and indirect effects on zooplankton. Ecology 74: 303-319. S-Plus (1995). S-Plus User's guide Pages. Saether, O. A. (1970). Nearctic and palearctic Chaoborus (Diptera: Chaoboridae), Fisheries Research Board of Canada Pages. Scheffer, M. (1989). Alternative stable states in eutrophic shallow freshwater ecosystems, a minimal model. Hydrobiological Bulletin 23: 73-83. Scheffer, M. (1990). Multiplicity of stable states in freshwater systems. Hydrobiologia 200/201: 475-486. Scheffer, M. (1991a). Fish and nutrients interplay determines algal biomass: a minimal model. Oikos 62: 271-282. Scheffer, M. (1998). Ecology of shallow lakes. New York, Chapman and Hall Pages. Scheffer, M., J. M. Baveco, D. L. DeAngelis, E. H. R. R. Lammens and B. Shuter (1995). Stunted growth and stepwise die-off in animal cohorts. The American Naturalist 145: 376-388. Schneider, D. C. (1994). Quantitative ecology. Spatial and temporal scaling. San Diego, Academic Press Pages. 192 Sinclair, A. R. E. (1995). Equilibria in plant-herbivore interactions. Dynamics, management and conservation of an ecosystem. A. R. E. Sinclair and P. Arcese. Chicago, The University of Chicago Press Pp. 91-113. Smyly, W. J. P. (1980). Food and feeding of aquatic larvae of the midge Chaoborus flavicans (Meigen) (Diptera: chaoboridae) in the laboratory. Hydrobiologia 70: 179-188. Stemberger, R. S. and J. J. Gilbert (1985). Body size, food concentration, and population growth in planktonic rotifers. Ecology 66: 1151-1159. Stenson, J. A. E. (1990). Creating conditions for changes in prey community structure by Chaoborus spa. in a lake in Sweden. Hydrobiologia 198: 205-214. Sutherland, J. P. (1981). The fouling community at Beaufort, North Carolina. The American Naturalist 118: 499-519. Sutherland, J. P. (1990). Perturbations, resistance, and alternative views of the existence of multiple stable points in nature. The American Naturalist 136: 270-275. Swift, M. C. (1976). Energetics of vertical migration in Chaoborus trivittatus larvae. Ecology 57: 900-914. Teraguchi, M. and T. G. Northcote (1966). Vertical distribution and migration of Chaoborus flavicans in Corbett Lake, British Columbia. Limnology and Oceanography 11: 164-176. Thompson, S. K. (1992). Sampling. New York, John Wiley & Sons, Inc. Pages. Vinyard, G. L. and R. A. Menger (1980). Chaoborus americanus predation on various zooplankters; functional response and behavioral observations. Oecologia 45: 90-93. Walters, C. J., E. Krause, W. E. Neill and T. G. Northcote (1987). Equilibrium models for seasonal dynamics of plankton biomass in four oligotrophic lakes. Canadian Journal of Fisheries and Aquatic Sciences 44: 1002-1017. Walton, W. E. (1988). The influence of zooplankton defenses on prey selection by young Chaoborus larvae. Journal of Plankton Research 10: 101-114. Walz, N. (1995). Rotifer populations in plankton communities: energetics and life history strategies. Experientia 51: 437-453. 193 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0074832/manifest

Comment

Related Items