Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Grazer control of bacterial abundance in a freshwater pond community Macinnis, Maura Jan 1997

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1997-0431.pdf [ 5.67MB ]
Metadata
JSON: 831-1.0074817.json
JSON-LD: 831-1.0074817-ld.json
RDF/XML (Pretty): 831-1.0074817-rdf.xml
RDF/JSON: 831-1.0074817-rdf.json
Turtle: 831-1.0074817-turtle.txt
N-Triples: 831-1.0074817-rdf-ntriples.txt
Original Record: 831-1.0074817-source.json
Full Text
831-1.0074817-fulltext.txt
Citation
831-1.0074817.ris

Full Text

GRAZER CONTROL OF B A C T E R I A L A B U N D A N C E IN A FRESHWATER POND C O M M U N I T Y by M A U R A JAN MACINNIS B.Sc, Mount Allison University, 1992 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE IN THE F A C U L T Y OF G R A D U A T E STUDIES (Department of Zoology) We accept thk^esis /^s3^iforming to the required standard THE UNIVERSITY OF BRITISH C O L U M B I A August 1997 © Maura Jan Maclnnis, 1997 . In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I 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 ^TPnVj^ The University of British Columbia Vancouver, Canada Date ^ / O V ^ DE-6 (2/88) Abstract Metazoan grazing of bacteria represents a potential pathway for the transfer of bacterial production to higher trophic levels. Freshwater cladocerans of the genus Daphnia are able to reduce bacterial abundance in lakes, but many experiments have been restricted to the summer months. It is therefore necessary to test the generality of Daphnia'?, role across seasons and across a broad range of food web configurations. The impact of mechanical grazing inhibition of Daphnid bacterivory, a potential outcome of algal blooms, has also not been addressed. I tested the ability of each of three crustacean zooplankton species, Daphnia pulex Leydig, Bosmina longirostris (O.F.M.), and Skistodiaptomus oregonensis (Lillj.), and a mixed rotifer community, to control bacterial abundance in 80 litre enclosures suspended in a freshwater pond. Experiments were conducted in August 1995 and some treatments were repeated in a second experiment in October 1995 to test for seasonal differences in grazer impact. Bacterial cell abundances at the end of the summer experiment were found to be significantly lower in Daphnia enclosures (1.87 x 106 cells ml"1) than in B. longirostris (3.91 x 106 cells ml"') and S. oregonensis (4.69 x 106 cells ml"1) enclosures, using repeated measures A N O V A . Bacterial abundances were also low in the absence of macrozooplankton in both summer (2.47 x 106 cells ml"1) and fall (1.19 x 106 cells ml"1). In contrast to the results observed in summer, Daphnia enclosures sustained high bacterial abundances in the fall. Daphnid grazing of bacteria appears to have been influenced by seasonal shifts in algae composition. The presence of a bloom of Elakatothrix sp. coincided with significantly higher bacterial cells abundances in Daphnia enclosures (2.76 x 106 cells ml"1 and 3.65 x 106 cells ml"1), while in both seasons, grazing by Daphnia reduced rotifer and ciliate abundances. Daphnid grazing of bacteria appears to be more susceptible to changes in grazing behaviour than other components of the food web. Thus, the presence of Daphnia can be expected to have a detectable effect on bacterial abundance, but the direction of impact may differ seasonally as algal composition changes. Smaller zooplankton are not able to reduce bacterial abundance, but the absence of macrozooplankton can also result in low bacterial abundances, due to the loss of indirect influences of macrozooplankton on the microbial food web. An experiment conducted to determine the impact of suspended particles on Daphnid grazing of bacteria resulted in an increase in bacterial abundance when grazing was inhibited by the presence of glass fibre filaments. The filaments, also resulted in a modest increase of bacterial abundance in the absence of macrozooplankton grazers. Mechanical interference with Daphnia grazing may mitigate Daphnia's potential for top-down control of the microbial food web. Suspended inorganic filaments were able to increase bacterial cell abundances in the absence of any substrate additions, indicating that the increased spatial heterogeneity and complexity afforded by suspended particles can cause a detectable enhancement of the microbial food web. Table of Contents Page Abstract i i Table of Contents iv List of Tables vii List of Figures xiv Acknowledgements xxii 1. General Introduction 1 1.1 Trophic interactions in microbial food webs 2 1.2 Other studies 7 1.3 Daphnia vs. small zooplankton in microbial food webs 8 1.4 Rationale and design 10 1.5 Zooplankton 14 1.54 Daphnia 14 1.52 Bosmina 17 1.53 Copepods 19 1.54 Rotifers 20 1.55 Predictions 22 2. Methods 24 2.1 Design 24 2.2 The G.G.E. Scudder Experimental Ponds 26 2.3 Enclosures 27 2.4 Preparation of treatments 31 2.4-1 Daphnia 31 2.4-2 Bosmina 31 2.4-3 Copepods 33 2.4-4 Rotifers 34 2.5 Collection of samples 34 2.6 Processing of samples 36 2.6-1 Zooplankton 36 2.6-2 Filtering capacity 40 2.6-3 Bacteria 40 2.6-4 Ciliates 41 2.6-5 Glass fibre filaments and green algae 42 2.7 Statistics 44 3. Results Part I: Persistence of treatments 46 3.1 Daphnia 46 3.2 Daphnia and filaments 57 3.3 Bosmina 58 3.4 Copepods 61 3.5 Rotifers 65 3.6 Other zooplankton 68 3.7 Temperature 70 4. Results Part II: Numerical responses to grazer manipulation 71 4.1 Bacteria abundance 71 4.1-1 Bacteria abundance- Summer 71 4.1-2 Bacteria abundance- Fall 78 4.1-3 Bacteria abundance- Seasonal comparison of DAPHNIA, BOSMINA, COPEPOD and ROTIFER 81 4.2 Response of rotifers to treatments 85 4.3 Response of ciliates 90 4.4 Response of algae 96 4.5 Filament effects- A fe-analysis 101 5. Discussion 106 5.1 Zooplankton Biomass 106 5.2 Effectiveness of treatments 107 5.3 Daphnia-rotifer interactions 109 5.4 Bosmina-rotifer interactions 111 5.5 Productivity and nutrient cycling 111 5.6 Daphnia and bacteria in Summer 112 5.7 Daphnia and bacteria in Fall 113 5.8 Other zooplankton and bacteria 115 5.9 Zooplankton-ciliate interactions 116 5.10 Algal blooms and grazing interference 118 5.11 Comparison to other studies 121 References 123 Appendix 1 135 Appendix 2 138 vi List of Tables Page Table 1. Important characteristics of systems dominated by Daphnia versus those dominated by smaller zooplankton, as observed in temperate, eutrophic lakes. Modified from Jurgens 1994. Table 2. A summary of the predictions regarding food web parameters for treatments in the Summer and Fall experiments. Note that some treatments were not performed in- both seasons. Table 3. Initial treatments added to enclosures 48 hours after filling of the bags with 54 um filtered water. Initial stocking densities are given in Table 4a, b, and c. Table 4a. Initial Daphnia sizes and filtering rated for both experiments. Table 4b. Initial Bosmina size and estimated clearance rates in the Summer experiment. Table 4c. Initial copepod density and estimated clearance rates in the Summer experiment. 23 25 32 32 32 vii Table 5. Length-weight regressions used in the calculation of 37 zooplankton biomasses. Equations used are of the form In W = In a + b In L, where L = length and W = mass of the zooplankter. Masses were calculated for each individual in a sample, and the total used to estimate zooplankton biomass for that sample. Table 6. Literature values of rotifer mass used in zooplankton biomass 38 calculations. Table 7a. Filtering rate equations used to calculate filtering capacity of 39 Daphnia pulex populations in Summer and Fall experimental enclosures. Filtering rates are expressed as ml individual"' h"'. L = length (mm). Table 7b. Filtering rate equations used to calculate filtering capacity of 39 Bosmina longirostris populations in Summer enclosures. Filtering rates are expressed as ml individual"1 h"1. L = length (mm). Table 7c. Per capita clearance rates of S. oregonensis, K. cochlearis, and 39 P. vulgaris used to estimate community filtering rates in experimental enclosures. viii Table 8a. A list of samples (indicating number of replicates) counted for each treatment in the Summer experiment. Blank cells indicate that no samples were counted on that date. Samples in bold text were included in the repeated measure A N O V A . Table 8b. A list of samples counted for each treatment in the Fall experiment. One sample was counted from each enclosure, for a total of three replicates per treatment. No samples were omitted or lost due to accident. Table 9. Range of bacterial abundance estimates observed during the two experiments in this study. A comparison with a range of literature values is shown in Figure 18. Values are given as cells ml"'. Table 10. Tukey multiple comparison test results following repeated measures A N O V A of natural log-transformed bacteria abundances (5 dates) in the Summer experiment. No comparison before August 17th (August 9, August 12) were significant (not shown). Significant differences are indicated in bold type, a= 0.05. 43 43 73 76 ix Table 11. Tukey multiple comparison test results based on repeated measure A N O V A of natural log-transformed bacterial abundances ( 5 dates) in the Fall experiment. No comparisons before November 2 n d (October 18th, October 25 t h, October 30 t h ) were significant (not shown). Significant differences at a = 0.05 are indicated in bold type. Table 12a. A N O V A comparison of natural log-transformed rotifer abundance on the final day of sampling. Includes all treatments from both the Summer and Fall experiments, a = 0.05. Table 12b. Levene's test of equality of error variances: tests the null hypothesis that the error variance of the natural log-transformed rotifer abundance is equal across treatments. The dependent variable is the natural log- transformed rotifer abundance on the final day of sampling. Treatments from both the Summer and Fall experiments were included. Table 13. A N O V A of natural log-transformed Elakatothrix sp. density on the final day of the Fall experiment. 80 87 87 99 x Table 14a. Repeated measures A N O V A of natural log-transformed bacterial abundance, comparing the DAPHNIA, DAPHNIA+F, FILAMENT, and ROTIFER treatments from the Fall experiment, 5 dates, a = 0.05. Table 14b. Repeated measures A N O V A of natural log-transformed bacteria abundance, comparing the DAPHNIA, DAPHNIA+F, FILAMENT , and ROTIFER treatments from the Fall experiment, 5 dates, a = 0.05. Table 14c. Levene's test of equality of error variances: tests the null hypothesis that the error variance of the natural log-transformed bacterial abundance in the enclosures is equal for the DAPHNIA, DAPHNIA+F, FILAMENT , and ROTIFER treatments. Table 14d. Tukey HSD multiple comparison tests between the bacterial abundances in the DAPHNIA, DAPHNIA+F, FILAMENT , and ROTIFER enclosures in the Fall experiment. Repeated measures A N O V A results are found in Table 14 a-c. Bacterial abundances have been natural log-transformed. Treatment comparisons which were also significant in an analysis which included the Fall COPEPOD treatment are shown in italic type. 103 103 103 104 xi Appendix 1 Repeated measures A N O V A of natural log-transformed 135 Table A bacterial abundance in Summer enclosures (5 dates). a= 0.05. Appendix 1 Repeated measures A N O V A of natural log-transformed 135 Table B bacterial abundance in Summer enclosures (5 dates). a= 0.05. Appendix 1 Levene's test of equality of error variances: tests the null 135 Table C hypothesis that the error variance of the natural log-transformed bacterial abundance in the Summer enclosures is equal across sampling dates. Appendix 1 Repeated measures A N O V A of natural log-transformed 136 Table D bacterial abundance in Fall enclosures (5 dates). a= 0.05. Appendix 1 Repeated measures A N O V A of natural log-transformed 136 Table E bacterial abundance in Fall enclosures (5 dates). a= 0.05. Appendix 1 Levene's test of equality of error variances: tests the null 136 Table F hypothesis that the error variance of the natural log-transformed bacterial abundance in the Fall enclosures is equal across sampling dates. X l l Appendix 1 Tukey multiple comparison test of natural log-transformed 137 Table G. rotifer abundance in enclosures. A l l treatments from both experiments were compared using A N O V A ; significance levels are indicated in the table, with significance at a = 0.05 given in bold type. xiii List of Figures Figure 1. Relative filtering rate of cladocerans for different particle 3 sizes. The fine line indicates that filtering rates on large protozoa do not conform to the model. Figure 2. The microbial and classical lake food webs. 4 Figure 3. Placement of treatments in enclosures in the Summer 28 experiment. The "sac" enclosures were sampled for zooplankton midway through the experiment. X denotes an unused enclosure. Figure 4a. Summary of results for the Summer experiment. 47 Figure 4b. Summary of results for the Fall experiment. 48 Figure 5a. Final zooplankton abundance in Summer enclosures. Values 49 given are the mean of 3 replicates. Error bars indicate 1 standard error. Figure 5b. Final zooplankton abundance in Fall enclosures. Values 49 given are the mean of 3 replicates. Error bars indicate 1 standard error. xiv Figure 6a. Zooplankton biomass on the final day of the Summer 50 experiment. Daphnia, Bosmina, and other zooplankton biomass were calculated using length-weight regressions (Table 5); rotifer biomasses were estimated using literature values. Error bars indicate standard error. Figure 6b. Zooplankton biomass on the final day of the Fall experiment. 50 Daphnia, Bosmina, and other zooplankton biomasses were calculated using length-weight regressions (Table 5); rotifer biomasses were estimated using literature values (Table 6). Figure 7. Abundance and biomass of Daphnia pulex in the 3 treatments 52 in both experiments to which Daphnia were added. Error bars indicate 1 standard error. Figure 8. Summer Daphnia length and weight distributions for 53 individuals counted (samples pooled for all three enclosures sampled). Figure 9. Length and weight distributions of Daphnia in the Fall 54 DAPHNIA samples (all three enclosures pooled). xv Figure 10. Estimated filtering capacities of Daphnia populations in 56 experimental enclosures, expressed as the time required for the zooplankton to clear the entire enclosure volume of bacteria and algae. Filtering capacity is estimated from the equations of Haney 1985 and Petersen et al. 1978. Error bars represent 1 standard error. Figure 11. Length and weight of Daphnia counted in final zooplankton 59 sampled from the three DAPHNIA+F enclosures (pooled). Figure 12. Estimated filtering capacities of Bosmina and Daphnia 60 populations in experimental enclosures expressed as the time required for the zooplankton to clear the entire enclosure volume of bacteria and algae. Filtering capacity is estimated from the equations of Haney 1985, Petersen et al. 1978, and DeMott 1982. Error bars represent 1 standard error. Figure 13a. Life stage distribution of copepods in enclosures at the end of 62 the Summer experiment. Values given are the mean of 3 replicate enclosures (2 in the BOSMINA treatment). Error bars indicate 1 standard error. Figure 13b. Life stage distribution of copepods in enclosures at the end of 62 the Fall experiment. Values given are the mean of 3 replicate enclosures. Error bars indicate 1 standard error. xvi Figure 14. Estimated time required by the copepod population to filter all 64 of the' enclosure volume, using flagellates and ciliates as reference prey items. Error bars indicate 1 standard error. Figure 15a. Rotifers at the midpoint in the Fall experiment. The animals 66 were added to the ROTIFER enclosures (3 replicates) at densities of 1200 individuals per litre, while the Killed Rotifer enclosures (2 replicates) received heat-killed inoculum. Error bars indicate one standard error. Figure 15b. Rotifer densities in the Fall experiment. The animals were 66 added to the ROTIFER enclosures at densities of 1200 individuals per litre; rotifers were initially reduced by filtering, while the Killed Rotifer enclosures (2 replicates) received heat-killed rotifer inoculum. Error bars indicate 1 standard error. Figure 16a. Noon temperatures measured in the Summer experiment (prior 69 to daily sampling). Figure 16b. Noon temperatures measures in the Fall experiment (prior to 69 daily sampling). Figure 17. Bacterial abundance measurements. Boxplots of experiments 72 and literature values. Estimates from each experiment are pooled. xvii Figure 18a. Bacterial abundance in Summer DAPHNIA enclosures. 74 Figure 18b. Bacterial abundance in Summer BOSMINA enclosures. 74 Figure 18c. Bacterial abundance in Summer COPEPOD enclosures. 74 Figure 18d. Bacterial abundance in Summer ROTIFER enclosures. 74 Figure 19. Mean of natural log-transformed bacteria abundance for the 77 DAPHNIA, BOSMINA, ROTIFER, and COPEPOD treatments in the Summer experiment. Homogeneous subsets are indicated by ellipses. Figure 20a. Bacterial abundance in Fall DAPHNIA enclosures. 79 Figure 20b. Bacterial abundance in DAPHNIA+F enclosures. 79 Figure 20c. Bacterial abundance in Fall COPEPOD enclosures. 79 Figure 20d. Bacterial abundance in Fall ROTIFER enclosures. 79 Figure 20e. Bacterial abundance in FILAMENT enclosures. 79 Figure 21. Mean of natural log-transformed bacterial abundance for the 82 DAPHNIA, DAPHNIA+F, COPEPOD , and ROTIFER treatments in the Fall experiment. Homogeneous subsets as determined by Tukey multiple comparisons are given by ellipses. See Appendix 1 Table G for significance levels. xvm Figure 22. A comparison of natural log-transformed bacterial abundances 83 in the Summer and Fall DAPHNIA treatments. The Summer experiment lasted 16 days, the Fall experiment 18. Data points represent the mean of three replicates. Figure 23. A comparison of natural log-transformed bacterial abundances 83 in the Summer and Fall ROTIFER treatments. The Summer experiment lasted 16 days, the Fall experiment 18. Data point represent the mean of three replicates. Figure 24. A comparison of natural log-transformed bacterial abundances 84 in the Summer and Fall COPEPOD treatments. The Summer experiment lasted 16 days, the Fall experiment 18. Data point represent the mean of three replicates. Figure 25a. Total rotifer abundance on the final day of the Summer 88 experiment (all species). Abundance values are given as a mean of 3 replicates (2 replicates in the BOSMINA treatment). Figure 25b. Total rotifer abundance on the final day of the Fall experiment 88 (all species). Abundance values are given as a mean of 3 replicates per treatment. xix Figure 26. Natural log-transformed rotifer abundance on the final day of 89 sampling both experiments. Homogeneous subsets (Tukey multiple comparison, Appendix 1 Table G) are indicated by the solid lines. Figure 27. Abundance estimates of the most common rotifer species in 91 both the Summer and Fall experiments. Values given are the mean of 3 replicates for each treatment (2 in the BOSMINA treatment). Figure 28a. Ciliate abundance in the DAPHNIA, BOSMINA, ROTIFER, and 92 COPEPOD treatments on the final day of the Summer experiment. Densities are given are means of 3 replicates (2 in the BOSMINA treatment). Error bars indicate 1 standard error. Figure 28b. Ciliate abundance in the DAPHNIA, DAPHNIA+F, FILAMENT , 92 ROTIFER, and COPEPOD treatments on the final day of the Summer experiment. Densities are given are means of 3 replicates (2 in the BOSMINA treatment). Error bars indicate 1 standard error. Figure 29. Natural log-transformed ciliate abundances on the final day of 93 the Summer and Fall experiments. Homogeneous subsets (Tukey multiple comparison) are shown by the black bars. xx Figure 30a. Species composition of the ciliate community in the Summer 94 enclosures. Densities are given as the mean of 3 replicate enclosures (2 in BOSMINA). Identification is made to the Order level except where indicated. Figure 30b. Species composition of the ciliate community in the Fall 94 enclosures. Densities are given as the mean of 3 replicate enclosures. Identification is made to the Order level except where indicated. Figure 31. Cell numbers of Elakatothrix sp. in the Fall enclosures. The 98 densities given for the DAPHNIA, DAPHNIA+F, ROTIFER, COPEPOD, and FILAMENT treatments are mean values for 3 replicate enclosures. Error bars represent 1 standard error. Figure 32. Natural log-transformed Elakatothrix densities in the Fall 100 experiment. Homogeneous subsets according to the Tukey multiple comparison procedure are indicated. The Daphnia-Rotifer comparison is significantly different at a = 0.053, all other comparisons are significant at a <0.05. xxi A ckno wledgements In completing this work, and in the efforts that preceded it, I am grateful for the help and support of many people. First and foremost, my supervisor, Dr. W. E. Neill, has been a mentor through many ups and downs. Through all the scientific debates, equipment follies and endless counting of zooplankton into the wee small hours, his attention and dedication as a supervisor and collaborator never wavered. Under his guidance, I have learned both the practice and the teaching of science, and for both those skills I will always be grateful. My lab mates in the Neill lab, Chris Schell, Jordan Rosenfeld, Chantal Ouimet, Karl Mallory and Heather Ferguson were always ready to get their feet wet for the cause of science. Danusia Dolecki spend many hours assisting me with the more inscrutable aspects of algal taxonomy. Wes Hochachka provided invaluable statistical advice, and an even more valuable sense of humour. John Richardson also provided statistical help, and generous feedback on very short notice. Lance Bailey and Alistair Blachford nursed frequently ailing lab computers back to health. Judith Maclnnis designed the prototype for my database and tutored me in the use of the software. David Montagnes did much to inspire my interest in microbial food webs, and helped me navigate both the literature and the lab with greater ease. Sean Doherty taught me the technique of epifluorescence microscopy and was always available for advice and friendship. I would like to acknowledge the support of my committee members past and present, John Richardson, Paul J Harrison, and Ken Hall, and the many members of the Fisheries Centre and the Zoology Department who have been generous with their advice, not to mention their lab equipment. The scientific side of this thesis tells only half the story. Without the support of my friends and family, I would have gone insane long before the first draft was written. My parents, Roland and Judith Maclnnis, have been my closest friends and strongest supporters in both the emotional and philosophical aspects my graduate career. My sister Margot has been a font of wisdom in every crisis, and my brother Drew has always been able to fix my damned computer and at the same time make me laugh. Arne Mooers, in love and friendship, inspired my poetry, and cajoled me to press on when the late nights became overwhelming. My friends Jennifer Garrison and Victoria Campbell have given me much sound advice, both personal and professional. This thesis project was generously supported by graduate fellowships from NSERC and the Department of Graduate Studies, and also by the NSERC operating grant of W. E. Neill. Additional funds were provided by an offspring improvement grant from Roland and Judith Maclnnis. xxii 1. General introduction It has been nearly two decades since limnologists and oceanographers began the task of integrating the microbial food webs into the classical theories of aquatic food web structure and function (Azam et al. 1983). This surge of interest in aquatic microbial ecology swelled on the heels of several methodological innovations which permitted ecologists to measure the density (Porter and Feig 1980), productivity (Fuhrman and Azam 1982), and diversity (Fuhrman et al. 1992) of aquatic bacteria. The ability to quantify changes at the base of the heterotrophic microbial food web allowed researchers to study the ecology of microbes and their grazers in relation to autotrophic and heterotrophic production, and eventually to integrate the "microbial loop" into the algae-zooplankton-fish model of lake food webs. This has permitted more accurate estimation of carbon flows and nutrient cycling in aquatic ecosystems. The nature of the relationship between the microbial and classical food webs was the subject of much early controversy (Ducklow et al. 1986, Sherr et al. 1986), as it became clear that in some systems (see Geertz-Hansen et al. 1987, Jeppesen et al. 1992), the secondary productivity of the microbial food web could be channelled to the macrozooplankton and thus become available to fish (Stockner and Porter 1988). In freshwater systems, researchers focussed their efforts on the distinguishing characteristics of food webs with microbial "links" as opposed to those exhibiting microbial "sinks" for organic carbon (Porter et al. 1988, Stockner and Porter 1988). It rapidly became clear that the freshwater cladoceran Daphnia was a key determinant of the fate of bacteria production (Stockner and Porter 1988, Gude 1 1988, Pace et al. 1990, Christoffersen et al. 1993), though this conclusion was reached in spite of some contradictory evidence (Pace and Funke 1991). Daphnia has long been considered a "keystone" species in freshwater ecosystems (Stockner and Porter 1988). In the absence of fish or invertebrate predation, Daphnia are competitively superior to most small-bodied zooplankton due to their relatively non-selective feeding behaviour (Hall et al. 1976). The preferred particle size spectrum for Cladocera is shown in Figure 1 (Jurgens 1995 after Gliwicz 1980). Daphnia is capable of grazing a much wider array of available algal resources (including particles as large as 150 um) and can suppress microzooplankton by both interference competition and exploitative competition simultaneously (Wickham and Gilbert 1991, 1993). It is thus not surprising to find that Daphnia spp. are often the most quantitatively significant links between the classical and microbial food web in lakes where they occur. The presence or absence of Daphnia can determine the magnitude of energy and nutrient transfer between the microbial and algae-zooplankton-fish pathways of the larger lake food web. 1.1 Trophic interactions in microbial food webs The major components of the microbial and classical food webs are depicted in Figure 2. The term "classical food web" is used by microbial ecologists to refer to the pathways of the lake food web traditionally considered to be based on autotrophic production. Thus the autotrophic algae fix inorganic carbon through photosynthesis and take up inorganic nutrients. Algae also release DOC (dissolved organic carbon), which becomes part of the D O M (dissolved organic matter) pool shown in Figure 2. The algae are consumed by herbivorous 2 Figure 1. Relative filtering rate of cladocerans for different particle sizes. The fine line indicates that filtering rates on large protozoa do not conform to the model. (modified from Jurgens 1994, originally from Gliwicz 1980) 3 Figure 2. The microbial and classical lake food webs Piscivores Planktivorous fish, Invertebrates Dissolved organic matter (DOM) Nutrients: N and P Carbon from "sloppy feeding" Predation Nutrient recycling zooplankton which convert algal carbon into animal biomass, excrete particulate organic matter and nutrients, and release algal cell contents into the surrounding water through "sloppy feeding". The herbivorous zooplankton are consumed either by carnivorous zooplankton or planktivorous fish, both of which return organic matter to the D O M and P O M (particulate organic matter) pools through excretion. In some systems, planktivorous fish are consumed by piscivorous fish. In literature prior to the mid-seventies, a microbial decomposer fauna was recognized, but its ecological role was restricted to the remineralization of refractory carbon compounds in the D O M pool (and therefore nutrient cycling). Modern convention now characterizes the bacterioplankton as a component of the heterotrophic food web (i.e. secondary productivity as distinguished from primary productivity), in which bacteria compete actively with algae for limiting nutrients (Currie and Kalff 1984, Currie et al. 1986, Currie 1990). At the base of the microbial food web, bacteria utilize the D O M pool to produce their biomass. Most bacterial cells are less than 1 urn in length and are vulnerable to direct grazing by protists. Small heterotrophic flagellates in the nanoplankton size range (2 - 20 um ) are the major grazers of the bacterioplankton, but larger flagellates and ciliates may graze bacteria as well (Sanders et al. 1989). Flagellates and ciliates may also graze algae, and prey on each other. Some microbial predators, unlike metazoan zooplankton, are capable of grazing prey which are equal to or larger than their own body size. Some species of algae are mixotrophic, grazing bacteria in addition to photosynthesizing, and they can be important grazers of bacteria in some systems (Boraas et al. 1988, Porter 1988, Sanders et al. 1989). Collectively these various microbes and protista are termed the microbial food web. 5 Where the algae are concerned, it is clear that the microbial and classical food webs are not functionally distinct components of the whole lake food web. Many of the organisms traditionally called "algae" are in fact autotrophic protists which are the same size as their heterotrophic counterparts, while some are mixotrophs which cannot be conveniently classified in the traditional autotroph/heterotroph food web paradigm. On average, about 40% of primary production fluxes through the bacteria in the photic zone of lakes and oceans (Cole et al. 1988)1. If the flow of carbon from the bacterioplankton and heterotrophic protists to macrozooplankton is of low magnitude (i.e. if neither bacteria nor flagellates and ciliates are grazed substantially by zooplankton), heterotrophic production by the bacteria is respired without reaching higher trophic levels. In such situations, the dynamics of the microbial components of the lake food web are of lesser importance to those who wish to understand the dynamics of algal, zooplankton and fish populations. In any event, to understand the dynamics of the microbial food web, it is necessary to quantify the biomass, productivity, and interactive pathways of its components. For questions involving the quantitative importance of the microbial food web to the processes of the classical food web, answers are often found in the zooplankton, and a cladoceran of the genus Daphnia often proves to be the determining factor. 1 Bacte r ia l production is about 20% of primary production in the photic zone (30% on an areal basis), and bacteria have a growth efficiency estimated at 50%. A s a comparison, zooplankton production is about 12% of primary production (see Cole et al . 1988). 6 1.2 Other studies A number of whole-lake and enclosure studies have addressed, directly or indirectly, the influence of grazer community structure on bacterial abundance (Riemann and Sondergaard 1986, Geertz-Hansen et al. 1987, Jeppesen et al. 1992, Markosova and Jezek 1993, Jurgens et al. 1994a, Pace and Cole 1996, Sarnelle 1997). Small enclosures studies tend to be well replicated and often involve direct manipulation of zooplankton abundances (Brett et al. 1994, Jurgens et al. 1994a, Sarnelle 1997). Zooplankton communities in large enclosure experiments are usually unreplicated (Riemann and Sondergaard 1986, Geertz-Hansen et al. 1987, Markosova and Jezek 1993) and indirectly manipulated using the presence/absence of fish (Riemann and Sondergaard 1986, Geertz-Hansen et al. 1987, Jeppesen et al. 1992, Markosova and Jezek 1993, Pace and Cole 1996). Manual zooplankton removal/addition is also common (Brett et al. 1994, Jurgens et al. 1994a, Sarnelle 1997). In all studies except Jurgens et al. 1994a, the presence of Daphnia caused a decrease in bacterial abundance. Bacterial abundance was elevated in the presence of small zooplankton grazers, and two studies were successful in maintaining a metazoan grazer-free treatment where bacterial abundance was lower than that observed in the presence of Daphnia (Brett et al. 1994, Jurgens et al. 1994a). There have been two attempts to separate the impact of the various small zooplankton species in "no Daphnia" treatments (Brett et al. 1994, Jurgens et al. 1994a), though only the study of Brett et al. (1994) attempted single-species manipulations. In one study, Bosmina longirostris has been observed to stimulate bacterial production in contrast to Daphnid?, top-down control of biomass and production (Jeppesen et al. 1992). In another study, copepods 7 exerted top-down control on ciliate abundance but bacterial abundance remained the same as that found in enclosures with no metazoan grazers (Jurgens et al. 1994a). The presence of Daphnia leads to a decrease in cell size in studies where bacterial biovolumes were measured (Jeppesen et al. 1992, Jurgens et al. 1994a). 1.3 Daphnia vs. small zooplankton in microbial food webs A summary of the known and predicted effects of Daphnia vs. small zooplankton grazing in lake food webs is given in Table 1 (modified from Jurgens 1994). This model was developed from the various lines of evidence for the impact of Daphnia, and also small zooplankton, on both the microbial and classical food webs. Other versions of this model have been mentioned in the literature (Gude 1988, 1990, Stockner and Porter 1988). Its predictions have been validated to various degrees (Jurgens 1994). The food web features described under a "Daphnia dominant" grazer community are analogous to the conditions observed in the absence of planktivorous fish populations, where large zooplankton such as Daphnia are mostly free from predation pressure and can attain high population densities. A "small zooplankton dominant" community would typically be observed under heavy size-specific planktivory such as that imposed by planktivorous fish or large invertebrate predators. The trophic cascade hypothesis is implicit in this model, which essentially characterizes food webs under "top down" control (Carpenter et al. 1985). 8 Table 1. Important characteristics of systems dominated by Daphnia versus those dominated by smaller zooplankton, as observed in temperate, eutrophic lakes. Modified from Jurgens 1994. Food Web Component Dominance of Daphnia (planktivorous fish absent) Dominance of small zooplankton (planktivorous fish present) Phytoplankton low biomass high biomass and diversity high grazing losses top down control nutrient limitation bottom up control mixotrophy Zooplankton Daphnia small cladocerans (Bosmina), rotifers, copepods Protozoa low numbers and diversity high numbers and diversity numerous interactions bacterivorous, algivorous and mixotrophic species Bacteria moderate bacterial abundance and biomass low morphological diversity small cell sizes high ratio of bacterial to primary production high numbers and biomass high morphological diversity grazing resistant forms: filaments, aggregates, attached bacteria low ratio of bacterial to primary production Detritus low standing stock, rapid turnover high standing stock aggregates colonized by bacteria and protozoans Nutrients elevated levels of dissolved nutrients nutrients bound in biomass, dissolved pools exhausted 9 1.4 Rationale and design The experiments in this study were designed to further investigate the interaction pathways between zooplankton and the microbial loop. Its has been asserted that Daphnia grazing may directly decrease bacterial abundance in freshwater lakes, while also stimulating productivity of the remaining bacteria indirectly via nutrient recycling and the release of algal carbon due to grazing (Jurgens 1994). The use of abundance as a measure of bacterioplankton dynamics can be problematic, as changes in the relative abundance of metabolically active cells can be masked by the greater abundance of dormant cells (del Giorgio and Scarborough 1995). The grazing impact of Daphnia, however, is potentially large enough to have a measurable effect on bacterial cell abundances. The ability of a predator to control the biomass of prey is strong indicator of top-down control of the food web (Carpenter et al. 1985, Carpenter et al. 1987, McQueen et al. 1989, Psenner and Sommaruga 1992). I therefore wished to assess the ability of Daphnia to suppress bacterial abundance, and contrast this with the grazing impact of smaller zooplankton species not known to exert top-down influence on the microbial loop. In seeking to establish and quantify metazoan links to the microbial food web, the impacts of particular grazer species are difficult to study in isolation. Only rarely are grazer "monocultures" (other than Daphnia) assessed for grazing impact in open lake water enclosures (Brett et al. 1994). The species-specific impacts of non-Daphnid zooplankton (especially non-cladocerans) on microbial food webs are usually inferred from laboratory studies of grazing rates on bacteria, protists and algae (Porter et al. 1983, Bleiwas and Stokes 1985, DeBiase et al. 1990, Sanders and Wickham 1993). 10 The experiments conducted in this study were designed to tease apart the impacts of 3 zooplankton grazers {Daphnia pulex Leydig 1860, Bosmina longirostris (O. F. Muller), and Skistodiaptomus oregonensis Lilljeborg 1889) and a mixed rotifer community (Keratella cochlearis Bory de St. Vincent and Polyarthra c.f. vulgaris Carlin 1943) on the microbial food web in the water column of a freshwater pond. While single-species impacts on the microbial food web have been studied previously (Brett et al. 1994), the small zooplankton species I examined have not been tested in isolation for their ability to control bacterial abundance. My enclosures were larger than those often used for measuring short-term microbial responses (Brett et al. 1994, Sarnelle 1997), and as I felt that previous failures to detect Daphnid?, impact on bacterial abundance were the result of experimental time scales that were too short. The durations of my experiments were 16 and 19 days. Central to the model of zooplankton-microbial food web interactions tested in this study is the generality of a particular grazer's impact on the algae, protista, and bacteria in the food web. Most limnological experiments take place in the summer months, and data from early spring, late fall and winter are generally sparse. A number of food web parameters can alter Daphnid?, clearance rates and retention efficiency for bacteria (Lampert 1987a, Porter et al. 1983). Algal composition, algal density, nutrient availability and abiotic factors such as temperature and turbidity all vary seasonally, and all can affect the feeding behaviour of Daphnia (Lampert 1987a). While the influence of these factors on Daphnia grazing is acknowledged (Jurgens 1994), the consequences for the microbial loop have not been comprehensively investigated in situ. I chose to repeat experiments seasonally to test the 11 generality of Daphnia's impact. Therefore the Daphnia treatments included in the summer experiment were repeated in the autumn of the same year (1995). Top down control by the grazer implies an ability to graze the available algae and microbes to the extent that standing stocks of bacteria and primary producers are reduced. But selection pressures of zooplankton on their prey, as well as bottom-up changes in nutrient regimes and abiotic factors, can induce responses in the algae that render the flora less vulnerable to grazing. Noxious, unpalatable, indigestible or colonial algae may be favoured, which are unavailable to zooplankton grazers. Inedible and sometimes inhibitory species often come to dominate the flora in the presence of Daphnia (Lampert et al. 1986, Sommer et al. 1986). While the inhibitory effects of algal toxins on Daphnia have been extensively studied, less is known about mechanical interference of filamentous algae with filter-feeding zooplankton (Webster and Peters 1978, Lampert 1987b). Inorganic particles (eg. suspended sediments) have been shown to interfere with the grazing of zooplankton populations (Kirk and Gilbert 1990, Kirk 1991), and so the possibility remains strong that there is a mechanical component to algal interference with zooplankton grazing. A number of food web parameters, such as the composition and abundance of algae, and also the presence of other particles (detritus) can alter Daphnid?, clearance rates and retention efficiency for bacteria (Lampert 1987a, Porter et al. 1983). Therefore, in addition to seasonal replication of the main experiment in the Fall, the impact of "model filamentous algae" on Daphnia grazing on bacteria was assessed. In conjunction with this, the effect of (inorganic) filament addition on bacteria density was tested in the absence of grazing pressure. Bacterial growth is stimulated by the presence of surface 12 area for attachment, as exemplified by the well-recognized problem of wall growth in experimental enclosures. Bacterial attachment to organic particles is common in freshwater, with attached bacteria comprising about 3% of the total bacterioplankton abundance (Kirchman 1983). Attached bacteria form a greater percentage of the total population (though never more than 10%) in the summer and fall than at other times of the year (Kirchman 1983). Attached bacteria are larger and metabolically more active than the free living bacteria (Kirchman 1983, Simon 1987, Gude 1990). The relative susceptibility of particle-bound bacteria to metazoan grazing varies according to grazer species (Schoenberg and Maccubbin 1985). It is possible that the presence of filamentous particles could stimulate bacteria growth by providing increased surface area for attachment. Aggregated growth forms also provide bacteria with refuge from protistan grazers (Gude 1990). Senescent algal blooms enhance the microbial food web by releasing organic carbon, but in providing a physical matrix for bacterial attachment they may also contribute a "mechanical" enhancement of microbial growth. Such an effect would increase the enhancement of the microbial food web in the latter stages of filamentous algal blooms. It is therefore likely that glass fibre filaments will inhibit Daphnid?, grazing on all components of the microbial food web, and increase bacterial abundance by providing increased surface area for attachment and growth. 13 1.5 Zooplankton The zooplankton communities in this experiment were manipulated to include only one species of macrozooplankton. Microzooplankton (rotifers) were also manipulated and were present in all the experimental zooplankton communities. The individual zooplankters used in my experiments are common limnetic species that have been subject to investigations of their impact on classical lake food webs. In the case of Daphnia, much is already known about its relationship to the microbial component of lake food webs. For Bosmina, S. oregonensis and the rotifers K. cochlearis and P. vulgaris, studies of microbial food web interaction are less common, as often the smaller zooplankton are studied collectively where they co-occur. The feeding behaviours of the zooplankton employed in this study, and their potential impact on the microbial loop, are summarized below. 1.5-1 Daphnia The dominance of Daphnia in freshwater food webs is a direct result of its competitive superiority over smaller zooplankton in grazing the < 20 urn algal size fraction (Hall et al. 1976, Gliwicz 1990). Competitive superiority and vulnerability to predation are positively related in the Cladocera (Bengtsson 1987), and it is the interaction of these major factors which structure zooplankton communities. In the absence of fish predation, larger bodied Cladocera are often able to competitively exclude smaller zooplankton, though the controversy surrounding this issue has hardly been settled (Dodson 1974, Romanovsky 1985, Bengtsson 1987, Gliwicz and Lampert 1993). Daphnia pulex has been shown to suppress the density of Bosmina longirostris, copepod nauplii and rotifers (Vanni 1986). Daphnia's dominance as a 14 pelagic grazer has spawned numerous investigations of its uniquely prodigious grazing ability (see Lampert 1987a for an extensive review). Daphnia spp. are filter feeders, with specialized feeding limbs having fine meshes which are able to retain particles less than 1 um in diameter. Daphnia pulex, with a mean filter-mesh size of about 0.4 um (Brendelberger 1985) is able to retain the larger bacteria (~1 um), but its filtering efficiency on the more numerous smaller cells (<0.5 um) is poor (Brendelberger 1991). However, large bacteria have higher growth rates and are responsible for more bacterial productivity than the smaller cells (Sherr et al. 1992). Daphnia is morphologically able to selectively crop the metabolically more active fraction of the bacterioplankton. The upper size limit on Daphnid? ingestion capability is correlated with the animal's body size (maximum length of adult animals -3.5 mm for the largest Daphnia species). Juveniles have finer meshes than adults (Brendelberger 1991), and filter mesh size is a phenotypically plastic trait that is developmentally responsive to food levels experienced by neonates (Lampert 1994). The smaller filter meshes of juveniles allow them to be more efficient feeders on the smallest size fraction of the planktonic food spectrum (Brendelberger 1991). In general, Daphnia clearance rates are highest on algae below 20 um (Gliwicz 1980 in Jurgens 1994), but clearance rates on large, soft-bodied protozoa can also be relatively high (Jurgens 1994). Feeding rates are influenced by food concentration, temperature, light, oxygen and pH, with nanoplanktonic algae comprising the most preferred component of the food spectrum. Daphnia pulex is capable of adjusting its feeding behaviour to lower its intake of low quality food and increase its ingestion of preferred species. Daphnia cannot completely 15 avoid grazing unwanted algae, and has been shown to be fairly non-selective when offered simple mixtures of food (DeMott 1982). Daphnia cannot reject individual particles, as food collected on its filter screens is transported en masse along the food groove to the mouth. The entire food groove may be cleaned by a rejection movement of the post-abdominal claw, but edible algae are removed along with the undesirable items (Lampert 1987a). Adult Daphnia pulex has been shown to feed on bacteria with clearance rates between 0.23 and 1.05 ml ind"1 h"1 (Jurgens 1994). Some studies have reported Daphnids grazing of bacteria to be enhanced when algal density is low (Sanders et al. 1989, Jurgens et al. 1994b), while other investigators report that the presence of larger particles enhances the retention efficiency for bacteria (Porter et al. 1983, Urabe and Watanabe 1991). "Clogging" of the filter meshes with larger (edible) algae may reduce the effective mesh size of the filtering limbs, while very low algal abundance may promote an increase in filtering rate for Daphnia with a concomitant increase in feeding rate on bacteria. Daphnia preys upon most of the major components of microbial food webs. Daphnia populations are able to suppress heterotrophic nanoflagellate abundance (Gude 1988, Weisse 1991, Jurgens and Stolpe 1995) and Daphnia ambigua is able to grow and reproduce on a diet of heterotrophic flagellates (Sanders and Porter 1990). Ciliates alone are not sufficient food for Daphnia (DeBiase et al. 1990), but Daphnia are able to graze small ciliates with the same efficiency as algae (Sanders and Wickham 1993) and can suppress ciliate abundance (Jack and Gilbert 1994). Daphnia is thus able to graze both the bacteria and bacterivorous protists. When its population density is high, it can clear the water of almost all edible algae and protists (with the exception of filamentous or colonial algae). This well known phenomenon 16 has been termed the "clearwater phase" where it occurs seasonally in lakes (Lampert et al. 1986, Sommer et al. 1986). 1.5-2 Bosmina Bosmina longirostris is a small bodied cladoceran which is capable of dual feeding modes. Its thoracic limbs are modified for both large-particle capture and small-particle filtering (DeMott and Kerfoot 1982, Bleiwas and Stokes 1985). Bosmina shows a strong preference for algal prey items over bacteria-sized particles, and will stop feeding in a pure bacterial suspension. Preconditioning on bacterial food sources only increased its preference for algae in grazing experiments (DeMott 1982). Bosmina's feeding mode is fundamentally different from that of Daphnia, and for this reason it is a highly selective feeder capable of efficiently avoiding ingestion of undesirable items (Burns 1968 in DeMott 1982). Bosmina has a large advantage in ingestion rate per unit biomass over that of Daphnia at low food concentrations. However, Bosmina's clearance rate is very sensitive to changes in food concentration, and at higher food concentrations its weight-specific ingestion rate is similar to that of Daphnia (DeMott 1982). Though it prefers algae in the <20 um size range, it is able to collect the larger cells in this size class more quickly. Most probably small particles are collected by filtration while the larger algae are captured by grasping (Bleiwas and Stokes 1985). It has been speculated that Bosmina's continuous swimming behaviour may increase its encounter rate with preferred prey items, which it could search out and actively grasp (DeMott 1982). 17 Daphnia relies on passive filtering and rejection mechanisms to avoid ingesting low quality food, while Bosmina is able to actively select high quality particles. Thus in situations where algal concentrations are high but the edible fraction is small, Bosmina can coexist with Daphnia even in the absence of fish predation. Where Daphnia feeds with low selectivity, Bosmina undergoes dietary switching and can discriminate between individual species of algae (DeMott and Kerfoot 1982). While the differences in feeding mode between Daphnia and Bosmina predict more complicated competitive outcomes than those suggested by the size-efficiency hypothesis (Dodson 1974, Hall et al. 1976), Bosminds feeding modes dictate that its impact on the microbial food web must also be fundamentally different from that of Daphnia. Bosmina's dual feeding mode allows it to selectively feed on highly edible flagellated algae, particularly when these prey items are present at low densities (Demott and Kerfoot 1982). The population growth rate of Bosmina has been correlated with flagellate density (Demott and Kerfoot 1982), and it has been shown to prefer grazing on flagellated algal cells over non-flagellated algae (Bogdan and Gilbert 1982). Flagellated algae can be autotrophic or mixotrophic, and are usually categorized separately from the heterotrophic flagellates in the literature. This designation is an artificial one where crustacean zooplankton are concerned, as heterotrophic protists are equal in quality to autotrophs as food for zooplankton (DeBiase et al. 1990, Sanders and Porter 1990, Sanders and Wickham 1993, Sanders et al. 1994). The potential of Bosmina to graze heterotrophic flagellates has not been tested experimentally. However, as with algae, the suitability of individual flagellate species as food for Cladocera likely varies, and where edible heterotrophic flagellates are present, Bosmina has the potential 18 to feed on them. In addition to its avoidance of bacteria as a food, Bosmina could increase bacterial standing stock still further by grazing heterotrophic nanoflagellates, which are the main bacterial predators (Fenchel 1982). Bosmina has been reported to capture ciliates at rates higher than its clearance rates for phytoplankton (Jack and Gilbert 1993), and ciliates are also well-documented bacterial grazers (Weisse and Muller 1990, Muller et al. 1991). Higher bacterial abundances are predicted in the presence of Bosmina than in Daphnia-dominated communities. If Bosmina were to graze bacterial predators selectively, bacteria standing stocks would be further enhanced. 1.5-3 Copepods Calanoid copepods such as Skistodiaptomus oregonensis are known to be highly discriminant grazers of freshwater algae (Butler et al. 1989). Copepods are capable of passive filter feeding on small particles, but their predominant feeding mode involves the capture and ingestion of larger cells. Some species have been shown to prefer larger algae and flagellates over smaller cells. Diaptomid copepods detect their prey primarily by mechanoreception and select their food actively (DeMott and Watson 1991). S. oregonensis itself is capable of a high degree of taste discrimination in accepting or rejecting prey items and the cells are usually tasted at the mouth before rejection (Demott and Watson 1991). When offered flavoured beads coated in algal extract, it showed a preference for flavoured beads and could discriminate among the "flavours" of algal species. In contrast, Daphnia shows very little taste discrimination, while Bosmina showed a modest taste response (Kerfoot 19 and Kirk 1991). This is consistent with both cladocerans' abilities to feed selectively on algae and protists. Skistodiaptomus oregonensis could not grow and reproduce on a bacterial diet, and though it reproduced well on abundant algal food, it achieved a higher reproductive rate when fed a mixed diet of algae and ciliates (Sanders et al. 1996). High reproductive rates were also achieved on a diet of ciliates alone. S. oregonensis did not thrive when its algal diet was supplemented with a heterotrophic nanoflagellate known to be suitable as a food source for Daphnia (Sanders et al. 1996). It is likely that copepods will not exert direct control on bacterial biomass, though they may be able to influence it through their grazing impact on bacterivorous ciliates and larger flagellates. S. oregonensis has low filtering rates compared to Cladocera of similar body size (Knochel and Holtby 1986). However, if copepods graze very selectively and at high rates they may be able to influence ciliate community structure (Burns and Gilbert 1993), and hence exert control of the microbial food web indirectly through predation on microbial grazers. Copepods can be significant predators on ciliates in marine microbial food webs, and also in freshwater systems (Sanders et al. 1996, Burns and Gilbert 1993, Sanders and Wickham 1993). 1.5-4 Rotifers The position of rotifers relative to the microbial food web has been subject to some dispute. Though in some systems they constitute a large fraction of the zooplankton biomass, their impact on the trophic dynamics of lake food webs is seldom regarded as 20 consequential (Bogdan and Gilbert 1982). However, at high abundances they have been found to have higher grazing rates on nanoplankton than Crustacea (Sanders et al. 1994). Recent synthesis indicates that rotifers are unlikely to exert top-down control over the microbial food web, though they may be able to alter the species composition and size spectrum of its components (Arndt 1993). It is difficult to separate rotifers as one "compartment" of lake food webs, as their range of body sizes, usually 100 to 500 um in length, overlaps those of large microbial grazers and small Crustacea (Pennak 1989). In addition, they have a range of feeding modes which allows for bacterivory, herbivory, and raptorial or "grasping" capture of single cells (Pennak 1989, Bogdan et al. 1980, Bogdan and Gilbert 1987). Many are omnivorous filter feeders which will consume any potential food item that falls within their preferred size range (Arndt 1993). The most abundant rotifer species in the South Campus experimental pond were Keratella cochlearis and Polyarthra c.f vulgaris, which constituted most of the rotifer fauna added to (or inadvertently present in) experimental enclosures. Both are common limnetic rotifer species (Pennak 1989). K. cochlearis has demonstrated an ability to feed on bacteria (Bogdan and Gilbert 1987, Sanders et al. 1989), though it may also show some selectivity for algal over bacterial cells (Bogdan et al. 1980, Gilbert and Bogdan 1981, Bogdan and Gilbert 1982.) It is a filter feeder capable of concentrating small particles in its feeding current and ingesting them en masse (Bogdan et al. 1980; Starkweather 1980; Bogdan and Gilbert 1987). Polyarthra prefers food in the l-40um range, and feeds on single, flagellated cells (Gilbert and Bogdan 1981; Bogdan and Gilbert 1982, 1987; Arndt 1993). It is a much more selective 21 grazer than K. cochlearis (Gilbert and Bogdan 1981), and does not graze bacteria (Sanders et al. 1989). For both species, the presence of a flagellum on an algal cell facilitates capture of the cell by the rotifer, though only Polyarthra seems to require this feature (Gilbert and Bogdan 1981). K. cochlearis has been observed to grasp algal cells by the flagellum in order to facilitate ingestion (Pourriot 1977). Polyarthra has also been observed to feed on species of Bodo, a nanoflagellate genus which includes bacterivorous species (Buikema et al. 1978). 1.5-5 Predictions The predictions regarding bacterial abundance, as well as ciliate and rotifer densities, are summarized for each treatment in Table 2. 22 Table 2. A summary of the predictions regarding food web parameters for treatments in the Summer and Fall experiments. Note that some treatments were not performed in both seasons. Treatment Predictions for Summer Predictions for Fall DAPHNIA Daphnia pulex' low bacterial abundance suppression of ciliates and rotifers low bacterial abundance suppression of ciliates and rotifers BOSMINA Bosmina longirostris high bacterial abundance suppression of ciliates NA COPEPOD S. oregonensis high bacterial abundance suppression of ciliates high bacterial abundance suppression of ciliates ^\ ROTIFER K. cochlearislP. vulgaris high bacterial abundance high ciliate abundance highest rotifer abundance high bacterial abundance high ciliate abundance highest rotifer abundance DAPHNIA+F Daphnia pulex under grazing inhibition NA high bacterial abundance moderate ciliate abundance moderate rotifer abundance FILAMENT Suspended glass fibre filaments NA increase bacterial abundance 23 2. Methods 2.1 Design Two experiments were performed to elucidate the potential influence of zooplankton grazers on the microbial food web of a small pond. The experiments were designed to detect whether a particular grazer community (ideally consisting of a single grazer species) could control bacterial abundance. These experiments exposed microbial food webs to simplified, strongly manipulated grazer communities over a time scale which encompassed many generations of the bacterial prey populations. Experiment 1 took place from August 9 to August 24, 1995. Experiment 2 was run from October 19 until November 4, 1995. The second experiment was designed to repeat treatments from experiment 1 later in the season. I will refer to experiment 1 as "Summer" and experiment 2 as "Fall" when making seasonal comparisons of the results. Some treatments from the Summer experiment could not be run in the Fall, and therefore two new treatments were added to the Fall experiment. The basic structure of both experiments included five grazer treatments with three replicate enclosures for each treatment. Treatments were randomly assigned to enclosures (15 out of the 20 enclosures were "experimental"). The treatments applied are described in Table 3. In Summer, treatments were selected to represent "Daphnia-dominatcd" (DAPHNIA treatment) and "small zooplankton-dominated" communities (BOSMINA, COPEPOD and ROTIFER treatments). Prior to treatment addition, each enclosure contained a natural pond phytoplankton/microbial community from which metazoan grazers had been removed. In the 24 Table 3. Initial treatments added to enclosures 48 hours after filling of the bags with 54 um filtered water. Stocking densities are given in Table 4 a-c. # Season Treatment Description Number stocked per bag Biomass stocked per bag (mg dry weight) Mean weight of individuals stocked ( M g ± 1 standard error) Mean length of individuals stocked (mm) 1 Summer Daphnia Daphnia pulex adults (> 1 mm) 350 19 54.7 ± 1.4 2.47 2 Summer Bosmina Bosmina longirostris 9600 9.3 0.97 ± 0.074 0.33 3 Summer Rotifer Keratella cochlearis Polyarthra vulgaris copepod nauplii (not counted), 1200 0.051* 0.043* not measured 4 Summer Copepod Diaptomus oregonensis adults and copepodites 1920 19.2** 10 not measured 5 Summer No Grazer no zooplankton added 6 Fa l l Daphnia Daphnia pulex adults (> 1 mm) 950 17 17.8 ± 0 . 4 1.5 9 Fa l l Daphnia + F Daphnia pulex adults (<1 mm) glass fibre filaments 950 8 x 10 7 17 17.8 ± 0 . 4 1.5 7 Fa l l Filament glass fibre filaments no zooplankton added 8 x 10 7 8 Fa l l Rotifer Keratella cochlearis Polyarthra vulgaris 1200 0.051 + 0.043* not measured 1 0 Fa l l No Grazer no zooplankton added "not measured directly; calculated using estimated biomass per individual from literature values *not measured directly; estimated from average of literature values of biomass for both species ** based on biomass given in Dumont et al . 1975 for calanoid copepods; very l ike ly an overestimate 25 Fall, two new treatments were added to investigate the effects of mechanical interference on Daphnia grazing. There is no established protocol for determining the appropriate duration of an experiment for a container of a given size. On the basis of a pilot study conducted in early spring, I estimated that seven days would be required to detect any effects, and doubled this estimate as a safety margin. The container size chosen was based on the need to hand-sort sufficient zooplankton for three replicates of each treatment while minimizing the handling time for the organisms. 2.2 The G.G.E. Scudder Experimental Ponds A l l experiments were conducted in Pond 13 of the G.G.E. Scudder Experimental Ponds on the University of British Columbia campus. The ponds are located in a clearing adjacent to old second-growth BC temperate rainforest. There are 13 morphologically identical ponds in close proximity. These artificial ponds were constructed in 1990, each with dimensions of 23 m X 23 m, with the sides of the pond sloping at a 3:1 ratio to a maximum depth of 3 m (Schluter 1994). They have a natural bottom substrate covering a thick plastic liner (Schluter 1994). The initial substrate depth was 30 cm, consisting primarily of sand and Texada limestone, but the deposition of sediment has increased at the centre of the pond due to slumping from the littoral zone. The pH of Pond 13 is slightly alkaline, and remained at or near 8.5 during the time of my experiments. In 1991, the ponds had initially been stocked with zooplankton and macrophyte vegetation from Paxton Lake, a mesotrophic lake on Texada Island. Pond 13 has been 26 unmanipulated since that time and serves as one of the "control" ponds in a long term study of the impact of stickleback on zooplankton and benthic communities. In Summer and Fall 1995, Pond 13 exhibited a typical small-bodied zooplankton community. Its late summer zooplankton assemblage consisted primarily of the copepod Skistodiaptomus oregonensis along with the cladocerans Bosmina longirostris and Diaphanosoma brachyurum. The most abundant rotifer species were Keratella cochlearis and Polyarthra c.f. vulgaris. Pond 13 is Ashless, however it had high densities of Chaoborus sp. larvae in summer 1995 and experienced heavy invertebrate planktivory. There were no D. pulex in Pond 13 in the Summer or Fall of 1995, and Daphnia has made only sporadic appearances in the pond in recent years. The bottom of Pond 13 is obscured by a thick carpeting of macrophytes. The water remains clear throughout the year and the bottom vegetation is always visible. This is in contrast the "fish" ponds adjacent to it, which have frequent algal blooms that greatly reduce water clarity. Pond 13 has come to adequately represent a "natural" pond in terms of its limnetic plankton and benthic invertebrate communities. Its recent origin has resulted in a system with low species diversity that facilitates manipulation and monitoring of its components. 2.3 Enclosures The enclosures were built of 6 mil (0.15 mm) clear polyethylene sheeting, with a maximum capacity of 100 litres (see Figure 3). The polyethylene sheeting was washed thoroughly with phosphate-free soap to removed any binders or lubricants remaining from the 27 Figure 3. Placement of treatments in enclosures in the Summer experiment. The "sac" enclosures were sampled for zooplankton at the midway through the experiment. X denotes an unused enclosure. 1 m 28 manufacturing process. The enclosures were tied into wooden floating frames which separated the open enclosures from the pond surface by a height of 10 cm. The two floating wooden frames were anchored in the centre of Pond 13. The dimensions of the enclosure bags and the position of assigned treatments are shown in Figure 3. The pond water which was used to fill the enclosures, was collected with a battery-powered plankton pump at approximately 1 m depth in the centre of Pond 13. The water was filtered through a 54 pm plankton net into 70 litre plastic containers and mixed thoroughly. Each batch of filtrate was divided and distributed equally across all enclosures until the enclosures were filled to 80 litres in volume. This process ensured that initial conditions were nearly identical for all enclosures. In both experiments, a natural algal/microbial Pond 13 community was retained in the enclosures, while all macrozooplankton and most of the rotifers were removed by the filtration. Further reducing the filter mesh size to screen out all rotifers would also have screened out large protozoa and algae. This would have prevented a natural microbial community from developing. Large dinoflagellates (Ceratiutn sp.) and other large algal cells were also removed by the filtering; any further reduction in large algae was not desirable. The presence of rotifers was unavoidable in all experimental enclosures. A solution of potassium phosphate and potassium nitrate, in an atomic N:P ratio of 20:1, was added to filled enclosures, for a total phosphorus addition of 10 ug/L and a nitrogen addition of 90.4 ug/L. This was done to ensure that the enclosures would be able to develop sufficient algal biomass before zooplankton were added, as well as sustain algal growth throughout the experiment. The bags were allowed to stand for 48 h prior to macro-29 and microzooplankton additions to permit the algae and protista to recover numerically from the effects of pumping and filtering. Each enclosure opening was covered with nylon mesh window screening to reduce illumination in the enclosures. A l l organisms were necessarily restricted to the upper 1 m of the water column and therefore were unable to migrate in response light levels. Shading reduced the possibility that light levels in the enclosures would be harmful to the plankton. Enclosures were thoroughly mixed twice daily using a long plastic stirring rod with a small paddle at the tip. Care was taken to stir gently while bringing up water from the bottom of the bag and loosening settled detritus. The enclosures had very little natural turbulence, and the stirring protocol prevented algae and nutrients from "settling out" of the enclosure system. Shortly after the zooplankton addition in the Fall experiment, filaments were added to enclosures receiving the glass fibre filament treatment. The filament solution was prepared by sonicating Whatman GF/F and GF/C filters in distilled water, until the filters were completely dispersed. The filters had previously been ashed at 450 °C for 24 h (Brinch-Iversen and King 1990). The fibre solution was settled and the filament density estimated; the filament solution was then added in aliquots to the bags to achieve an initial density of 1000 filaments per ml. Throughout the experiment there was considerable loss of filaments due to settling despite the stirring protocol. To counteract this, additions of filament solution were given to the enclosures several times throughout the 18 day experimental run. Preliminary laboratory estimates indicated that all of the filaments would have settled out of solution after 24 h; stirring occurred every 12 h. Filament densities were therefore quite variable. Filament density in recently stirred enclosures was 4200 + 1300 ml"1 (mean + 1 standard error) on the final day of the Fall experiment. 30 2.4 Preparation of treatments 2.4-1 Daphnia Daphnia pulex used in experiments were collected from an ornamental pond north of Main Library, on the University of British Columbia campus. There are Daphnia present in "Library Pond" from early spring until late fall. They were easily collected in large numbers using a 10 X 20 cm aquarium net. Daphnia were sorted and counted in the lab; adults estimated to be larger than 1 mm were selected using a large-bore pipette. Any individuals exhibiting ephippia (resting egg cases) were excluded. The sorted stock was then sampled to obtain a size distribution for the experimental animals. Daphnia were added to the enclosures less than 24 h after collection, and were kept incubated in the dark at 16°C until the bags were stocked. Stocking densities were 5 individuals L"1 in experiment 1 and 12 individuals L"1 in experiment 2; the total biomass stocked per enclosure was 19 mg in Summer and 17 mg in Fall (Table 3). This biomass of Daphnia was chosen to give a final population filtering capacity (allowing for reproduction) of about 1/3 of the enclosure per day. D. pulex size in the source populations (August vs. October) differed substantially (Table 4a); stock densities were adjusted in Experiment 2 to maintain comparable Daphnia biomass between experiments (Table 3) . 2.4-2 Bosmina Bosmina longirostris added to experimental enclosures were collected from Pond 13 using a plankton net with a mesh size of 54 um, and sorted in the lab overnight prior to addition to the enclosures. Bosmina were separated by hand from the other plankton; this 31 Table 4a. Initial Daphnia sizes and filtering rates for both experiments. Month Mean Length (urn) Mean Weight (ug) Peterson1 (ml ind 1 d"1) Haney2 (ml ind"1 d"1) Number L" 1 August 2467 54.7 31.6 27.2 5 October 1523 17.8 14.4 10.6 12 1 calculated from the equation for hourly grazing rate on natural bacteria; Peterson et al. 1978 2 Haney 1985 Table 4b. Initial Bosmina size and estimated clearance rates in the Summer experiment Month Mean Length (urn) Mean Weight (Mg) Filtering rate on bacteria (ml ind"1 d"1) Filtering rate on flagellates (ml ind"1 d"1) Number L" 1 August 332 0.97 0.43 1.87 120 Table 4c. Initial copepod density and estimated clearance rates in the Summer experiment. Month Mean Weight ("g) Filtering rate on ciliates (ml ind"1 d"1) Filtering rate on flagellates (ml ind"1 d"1) Number L" 1 August 10* 7.68 4.8 20 weight is an estimate from literature (Dumont et al. 1975) 32 was facilitated by Bosmina's phototactic response and relatively rapid swimming speed. Initial stocking densities were targeted at 100 individuals per litre with a 20% allowance for handling mortality, therefore 120 individuals were added for each litre of enclosure volume. Assuming some eventual biomass increase, this density was roughly estimated to allow a population filtering capacity (L pop"1 d"1 ) of 1/3 of the enclosure volume. Animals were stocked in enclosures within 24 h of collection, after being kept in Nalgene carboys overnight in a dark incubator at 16°C. Initial Bosmina size and filtering capacity is given in Table 4b. 2.4-3 Copepods After most of the Bosmina had been removed from the collected plankton, the remaining Skistodiaptomus oregonensis could only be separated from the co-occurring Diaphanosoma brachyurum by allowing the concentrated animals to remain in the 20 L collection carboys overnight (in a dark incubator at 16°C). S. oregonensis survived this treatment, but D. brachyurum eventually collided with the walls of the container and adhered or were trapped at the air/water interface. Few Diaphanosoma were left alive after 24 h, and the remaining zooplankton in the carboy were almost exclusively S. oregonensis. Copepods were stocked at an initial density of 24 animals per litre of enclosure volume, which included a 20% allowance for handling mortality. As copepod filtering rates are much lower per unit biomass than those of Cladocera, an estimated copepod biomass equal to that of Daphnia was added (Table 4c). Biomass estimates were made using values given for calanoid copepods in Dumont et al. 1975. 33 2.4-4 Rotifers Rotifers were harvested from experimental ponds which had high rotifer densities at the time the experiments were conducted; source ponds were chosen based on their low densities of S. oregonensis nauplii, which could not be separated from the collected rotifers. The stock of rotifers added was combined from Ponds 3, 5, and 7, all of which contain limnetic or benthic stickleback. Rotifers were collected with a 54 um mesh-size plankton net, and the collected plankton was passed through a 118 um sieve to gently filter out all large zooplankton and most of the larger copepod nauplii. Small nauplii could not be separated from the rotifers by filtration. After the 118 pm filtration, rotifers showed some mortality due to sieving. Further reductions in mesh size increased handling mortality. To reduce the stress on the rotifers, they were harvested, concentrated, their abundance estimated, and the stock added to enclosures as rapidly as possible (within 1 h of collection). Initial stocking densities were 1,200 individuals per litre. Despite this precaution, some mortality of rotifers in the stocking carboys was observed. I did not attempt to equalize rotifer biomasses to that of Daphnia (200 ug L"1 ), though biomass estimates for K. cochlearis in Dumont et al. 1975 indicate that the biomass of rotifers added may have been as much as half the Daphnia biomass. 2.5 Collection of Samples Enclosures were sampled for bacteria and algae daily between noon and 14:00. Each bag was thoroughly stirred prior to sampling, and let stand for a few minutes to allow large detritus to settle. Three depth-integrated water samples were then taken from each bag using 34 a weighted length of polyethylene tubing (1 cm diameter). The subsamples from different areas of the bag were pooled (approx. 250 ml total volume). Any macrozooplankton captured were removed with a pipette and returned to the enclosure. The water sample was gently mixed and a 15 ml sub-sample for bacterial counts was taken and preserved with 2% glutaraldehyde. The remainder of the sample was preserved with acid Lugol's solution. Temperature was recorded daily2 in enclosure 1, initially with a probe accompanying the pH meter and later with a hand-held thermometer after the probe proved unreliable. pH was sampled with a portable pH meter (manually corrected for temperature) in all enclosures every 2 or 3 days3. Rotifers in the ROTIFER enclosures were sampled at the midpoint of each experiment (August 17, 1995 and October 25, 1995), to give an estimate of rotifer densities. Two additional enclosure bags were included in the Fall experiment which had been treated in the same fashion as the ROTIFER enclosures, but the inoculum of harvested rotifers was heat-killed before addition to the enclosures. These "killed rotifer" treatments were sampled for zooplankton at the same time as the other "rotifer" enclosures. The volume of the midpoint sample was 2 L from each enclosure. Midpoint macrozooplankton samples could not be taken from the experimental enclosures, though samples were taken from non-experimental enclosures included for this purpose. These samples were counted to determine whether added zooplankton survived until the midpoint of the experiment, and are not included in the presentation of results. Final rotifer and macrozooplankton samples were taken at the end of temperature samples were omitted on a few dates in the Summer experiment 3 a malfunctioning probe necessitated the exclusion of some Summer p H sampling 35 each experiment by emptying the entire contents of the enclosure through a 54 um plankton net and preserving all zooplankton in sugared formalin. 2.6 Processing of samples 2.6-1 Zooplankton Macrozooplankton samples were counted using a Wild M5 dissecting microscope equipped with a drawing tube, digitizer pad and associated microcomputer. Subsamples were taken using a plankton splitter and counted to give a minimum of 300 individuals of the main taxa present. With the exception of rotifers, all individual zooplankton counted were also measured for length. Macrozooplankton masses were calculated using the equations given in Table 5. Any individuals which could not be measured due to poor orientation or physical damage to the organism were assigned the mean weight for that taxon for that particular replicate. Mass for each individual was calculated from length-weight regressions and summed to give the total biomass for the sample. This obviates the need for any correction factors associated with the use of mean zooplankton length to estimate biomass for the sample (see McCauley 1984 for a review). Mean length is a more accurate measure when measuring only 30-50 animals in a sample, but as I measured many more individuals (about 300 per sample), summing individual calculated weights provides a better estimate. Rotifers were enumerated using the digitizer pad; accurate lengths could not be determined using the digitizer and microscope available; biomasses were calculated using the average species-specific biomass values available from the literature (Table 6). 36 Table 5. Length-weight regressions used in the calculation of zooplankton biomasses. Equations used are of the form In W = In a + b In L, where L= length and W = mass of the zooplankter. Masses were calculated for each individual in a sample, and the total used to estimate zooplankton biomass for that sample. Species Name a b units (length/weight) Source Daphnia pulex 0.00624 2.4 mm/mg Paloheimo et al. 1982 Diaphanosoma brachyurum 0 3.0468 mm/ug McCauley 1984 (Bottrell et al. 1976) Bosmina longirostris 0 2.5294 mm/ug McCauley 1984 (Bottrel et al. 1976) Chydorus sphaericus 0 3.636 mm/ug McCauley 1984 (Rosen 1981) Simocephalus vetulus 7.43 3.28 mm/ug Dumont 1975 Skistodiaptomus oregonensis nauplia 0 2.1547 mm/ug Malley et al. 1989 Skistodiaptomus oregonensis copepodite 0 2.4235 mm/ug Malley et al. 1989 Skistodiaptomus oregonensis adult 0 2.5384 mm/ug Malley et al. 1989 37 Table 6. Literature values of rotifer mass used in zooplankton biomass calculations. Species Name Weight (112 per ind) Reference Keratella cochlearis 0.0105 Ruttner-Kolisko 1977 in Malley et al. 1989 Keratella cochlearis 0.005 Hall et al. 1970 in Malley et al. 1989 Keratella cochlearis 0.0035 Nauwerk 1963 in Malley et al. 1989 Keratella cochlearis 0.005 Lewis 1979 in Malley et al. 1989 Keratella cochlearis 0.001 Berman et al. 1982 in Malley et al. 1989 Keratella cochlearis 0.07 Bottrell 1976 in Malley et al. 1989 Keratella cochlearis 0.11 Dumont et al. 1975 Keratella cochlearis 0.049 Schindler and Noven 1971 in Malley et al. 1989 Keratella cochlearis 0.07 Makarewicz and Likens 1979 in Malley et al. 1989 Keratella cochlearis 0.013 Comita 1972 in Malley et al. 19889 Keratella cochlearis 0.0337 mean value used in biomass calculations Keratella quadrata 0.35 Dumont et al. 1975 Keratella quadrata 0.32 Dumont et al. 1975 Keratella quadrata 0.335 mean value used in biomass calculations Lecane sp. 0.028 Malley et al. 1989 Lecane sp. 0.2 Bottrell et al. 1976 Lecane sp. 0.038 Malley et al. 1989 Lecane sp. 0.08867 mean value used in biomass calculations Polyarthra vulgaris 0.02 Lewis 1979 in Malley et al. 1989 Polyarthra vulgaris 0.098 Schindler and Noven 1971 in Malley et al. 1989 Polyarthra vulgaris 0.043 Doohan 1973 in Malley et al. 1989 Polyarthra vulgaris 0.06 Makarewicz and Likens 1979 in Malley et al. 1989 Polyarthra vulgaris 0.0385 Nauwerck 1963 in Malley et al. 1989 Polyarthra vulgaris 0.0519 mean value used in biomass calculations Synchaeta sp. 0.013 Malley et al. 1989 Synchaeta sp. 0.156 Malley et al. 1989 Synchaeta sp. 0.07 Malley et al. 1989 Synchaeta sp. 0.366 Malley et al. 1989 Synchaeta sp. 0.27 Dumont 1975 Synchaeta sp. 0.26 Dumont 1975 Kynrhnpta sp 0.1892 m p n n vnlnp nspri i n hinmact; c a l c u l a t i o n s 38 Table 7a. Filtering rate equations used to calculate filtering capacity of Daphnia pulex populations in Summer and Fall experimental enclosures. Filtering rates are expressed as ml individual1 h'1. L= length (mm). Source Equation r Temperature Food Item Time Peterson et al . 1977 F = 0.294 L 1 6 6 0.93 8 °C natural bacteria midnight Haney 1985 F = 4.467 L 2 I X ) 0.98 3.1 - 4.0 °C labelled yeast (tracer cells) night Table 7b. Filtering rate equations used to calculate filtering capacity of Bosmina longirostris populations in Summer enclosures. Filtering rates are expressed as ml individual"1 h"1. L= length (mm). Source Equation r 2 Temperature Food Item Time DeMot t 1982 F = 0.106 L ' 6 6 3 0.69 15 °C aerobacter night DeMot t 1982 F = .598 L 1 8 7 0.87 15°C chlamydomonas night Table 7c. Per capita clearance rates of S. oregonesis, K. cochlearis and P. vulgaris used to estimate community filtering rates in experimental enclosures. Source Species Prey Clearance rate Sanders and W i c k h a m 1993 S. oregonensis mixed ciliates <30um 0.32 m l ind" 1 h"1 Sanders and W i c k h a m 1993 S. oregonensis paraphysomonas 0.2 ml ind" 1 h"1 Sanders et al . 1994 rotifers 1 4 C-label led flagellate 0.051 ml ind"1 h 1 Bogdan et al . 1980 Polyarthra dolichoptera bacteria 0.01 u l ind" 1 h"1 Bogdan et al . 1980 Polyarthra dolichoptera chlamydomonas 1.69 ul ind" 1 h"1 Bogdan et al. 1980 Keratella cochlearis bacteria 0.29-0.46 u l ind" 1 h"1 Bogdan et al . 1980 Keratella cochlearis chlamydomonas 0.76-6.41 u l i n d 1 h"1 39 2.6-2 Filtering capacity Total filtering capacity of the Daphnia added to enclosures in both experiments was calculated from the length-weight regressions in Table 7a. Daily (or hourly) filtering rates for each individual animal in a sample were calculated and summed. This filtering capacity was expressed as total volume filtered by the Daphnia pulex population per day. The total volume of the enclosure was divided by this number to estimate the amount of time required by the Daphnia population in an enclosure hypothetically to filter all the water in the bag. Similar calculations were made for the BOSMINA, COPEPOD, and ROTIFER treatments. The equations used are summarized in Table 7b-c. Where possible, clearance rates for each species on bacteria and flagellates were calculated, but as the clearance rates for S. oregonensis were highest on ciliate and negligible on bacteria, ciliate and flagellate clearance rates were employed for this species. In the case of Bosmina, it was possible to use published regressions of clearance rate to body length to calculate population clearance rates (L population"1 day"1 ). For S. oregonensis and rotifers, only measured per capita rates were available. The mean filtering capacities of the Daphnia, Bosmina, S. oregonensis and rotifer (not reported) populations are expressed as the estimated time required for the population to filter the entire volume of the enclosure. 2.6-3 Bacteria Gluteraldehyde-preserved bacterial samples were refrigerated at 5°C immediately after collection for storage until processing. A two millilitre sub-sample was stained with 4, 6 diamidino-2-diphenylindole (DAPI) at a concentration of 5.8 ug/ml, and filtered under low 40 vacuum onto a 0.22 um black membrane filter (Millipore). The filter was mounted onto a glass slide and frozen until counting (Porter and Feig 1980). Slides were made from each sample within one month of collection. A l l samples were subject to the same storage time. Slides were viewed using a Nikon inverted microscope equipped for epifluorescence. High bacterial density in the samples precluded efficient direct counting using the epifluorescence microscope, as fading of the stain often occurred before all cells in the field view could be counted. This made it necessary to develop a counting method which would accurately and quickly record the entire field view for later counting. After trial tests to determine accuracy of image recording, samples to be counted were photographed using T M A X 400 film set at 5 s exposure time. This method allowed me to record the presence and shape of even weakly fluorescing cells. Bacterial counts were made directly from the film negative using a dissecting scope with 16X magnification. A 2 cm grid in the centre of each negative was examined and counted, as focus irregularities near the edges of the film made it undesirable to count the entire photograph. A minimum of 10 randomly chosen frames (approx 1000 cells) were counted for each sample. This is in accordance with other methods for bacteria counting (Kirchman 1993). The area counted was larger than the area usually counted with the microscope eyepiece graticule (grid) used in direct counting, and this increased the accuracy of the count. 41 2.6-4 Ciliates Lugol-preserved samples were settled in 25 ml counting chambers and enumerated for ciliates at 150X magnification. The entire area of the chamber was counted. Results were reported as total ciliate numbers per litre. Ciliates were identified to the order level (Pennak 1989), but as abundances of individual taxa were often too low to extrapolate densities for each species, only the total for all taxa could be reliably estimated. In addition, it is not advisable to attempt precise taxonomic identification for ciliates preserved in Lugol's. Protargol staining is generally required for accurate determination to the species level. 2.6-5 Glass Fibre Filaments and Green Algae Final densities of glass fibre filaments added to "FILAMENT" treatments were estimated from algae samples taken on the final day of the experiment. Subsamples of the Lugol-preserved algal samples were settled in 25 ml settling chambers and counted at 600X magnification. A bloom of gelatinous green algae, Elakatothrix sp. (Wille 1898), occurred in the second experiment. Its densities were estimated by counting a minimum of 300 cells (usually about 5 fields) at 600X magnification. 2.7 Statistics A l l means are reported + one standard error. Though bacteria were sampled daily, a subset of the samples was chosen and counted to give estimates of bacterial abundance throughout each experiment. In the Summer experiment, samples from the DAPHNIA, BOSMINA and NO GRAZER treatments were counted on 42 Table 8a. A list of samples (indicating number of replicates) counted for each treatment in the Summer experiment. Blank cells indicate that no samples were counted on that date. Samples in bold text were included in the repeated measures A N O V A . Date Experiment DAPHNIA BOSMINA COPEPOD (failed copepod) ROTIFER August 9 Summer 3 3 3 3 3 August 11 Summer 3 3 3 August 12 Summer 3 3 3 3 2* August 14 Summer 2* 3 3 August 17 Summer 3 3 3 3 3 August 21 Summer 3 3 2* August 23 Summer 3 3 3 3 3 August 24 Summer 3 3 3 3 3 replicate sample missing due to errors in processing. Table 8b. A list of samples counted for each treatment in the Fall experiment. One sample was counted from each enclosure, for a total of three replicates per treatment. No samples were omitted or lost due to accident. Date Experiment DAPHNIA DAPHNIA+F COPEPOD FILAMENT ROTIFER October 18 Fall 3 3 3 3 3 October 25 Fall 3 3 3 3 3 October 30 Fall 3 3 3 3 3 November 2 Fall 3 3 3 3 3 November 4 Fall 3 3 3 3 3 43 more dates than the ROTIFER and COPEPOD treatments. The samples counted for both experiments are listed in Table 8a-b. For both Summer and Fall experiments, the complete data set for statistical analysis consisted of 5 sample dates. Bacterial abundance measurements were natural log-transformed and examined statistically using a repeated measures A N O V A procedure with SPSS 7.5 statistics software. When several measurements are made on the same experimental unit (each enclosure on 5 sample dates), this procedure assumes a correlation of the measurements within the same enclosure and separates this from the total variation. It is therefore possible to partition the effects of time (the Date variable) from the measurements of treatment effects (the Treatment variable). Date and Date*Treatment interaction are within-subject effects. Each enclosure is therefore analogous to a "block" in a randomized complete block design. The between-subjects aspect of the analysis examines the variation due to "between enclosures" effects (i.e. the treatments applied). One limitation of this procedure is that a missing sample will result in the exclusion of the enclosure from the analysis. This was the case for enclosure 9 on August 12th, as the bacteria sample from this enclosure was damaged during processing. In order to avoid exclusion of the enclosure from the statistical analysis, the missing value was replaced with an estimate, which was generated using a linear regression of abundance measurements from enclosure 9. This aspect of the analysis is discussed in Appendix 2. For each date in the bacteria analysis, post-hoc comparisons were performed using the Tukey HSD procedure. For some dates in the repeated measures analysis, the homogeneity of variance assumption was violated. A N O V A is generally robust to violations of this assumption, but this not universally true (Kirk 1982, Winer et al 1991). No transformation 44 of the data was able to stabilize variance in these instances. Results are reported as obtained and the outcome of the Levene test for homogeneity of variance is given where significant results indicate that caution is warranted. I consider it unlikely that a small departure from a nominal significance level of 0.05 warrants concern that the observed treatment effects are a statistical artifact. 45 3. Results Part I: Persistence of treatments In order to determine if the desired treatments were successfully implemented, I examined and compared the abundance and biomass of the zooplankton in the enclosures at the end of each experiment. The following results are split into two main sections: (1) an evaluation of the "success" of the treatment implementation (did the experimental manipulation produce the desired zooplankton communities under comparison?) and (2) an evaluation of the effect of each treatment on the microbial food web (what was the impact of each zooplankton community?). Figure 4a and 4b summarize the organizational framework for evaluating and presenting results. The mean zooplankton densities for each treatment on the final day of each experiment are given in Figure 5a (Summer) and 5b (Fall). The biomass estimates for zooplankton in each treatment are given in Figure 6a (Summer) and 6b (Fall). The densities, biomass, and filtering capacities of the zooplankton treatments will be discussed below. In some treatments the final zooplankton species composition differed from the initial single-species treatment added to bags at the start of each experiment, and these outcomes are also noted below. 3.1 Daphnia Figures 6a and 6b illustrate the high zooplankton biomass in enclosures with added Daphnia relative to other treatments. As expected, final Daphnia abundances exceeded the stocking densities of 5 and 12 individuals L"1 in both the Summer and Fall treatments. Final 46 Figure 4a. Summary of results for the Summer experiment Daphnia Bosmina Rotifer Copepod No Grazer adult Daphnia added at 5 per L adult Bosmina added at 120 p e r L mixed rotifers added at 1200 p e r L adult copepods and copepodites added at 20 per L large zooplankton, copepod nauplii and most rotifers excluded Evaluate Efficacy of Zooplankton treatments: Did treatments behave as Results Part 1 intended? V Daphnia population increase in biomass Daphnia dominates yes Bosmina high adult population copepod increase biomass; in biomass moderate (slight) rotifer biomass; high rotifer Copepods density dominate yes v re-label as \ "Copepod low adult copepod biomass treatment omitted rotifer population increased; no macro zooplankton Rotifers dominate re-label as "Rotifer" Evaluate Response to Treatments: Results Part 2 I What is the effect on the microbial food web? low bacteria abundance low ciliate abundance suppression of rotifers high bacteria abundance moderate ciliate abundance high rotifer abundance low bacteria abundance high ciliate abundance moderate rotifer abundance high bacteria abundance low ciliate abundance moderate rotifer abundance Daphnia Bosmina Rotifer Copepod Figure 4B. Summary of results for the Fall experiment Daphnia adult Daphnia added at 12 per L Daphnia +F Filament adult Daphnia added at 12 per L glass fibre filaments added at 1000 per m l zooplankton, and most rotifers excluded -glass fibre filaments added at 1000 per m l Rotifer mixed rotifers added at 1200 p e r L No Grazer large zooplankton, copepod nauplii and most rotifers excluded Evaluate Efficacy of Zooplankton treatments: Results Part 1 Did treatments behave as intended? Daphnia population increase in biomass Daphnia dominates Daphnia population increase in biomass (less than in replicates without filaments rotifer population increased; no macro zooplankton Rotifers dominate low adult copepod biomass; copepods present; rotifers dominate rotifer population increased; no macro zooplankton Rotifers dominate yes yes yes re-label as "Copepod" re-label as "Rotifer" \ / Evaluate Response to Treatments: Results What is the affect on the micropbial Part 2 food web? T high bacteria abundance low ciliates rotifers suppressed bacteria abundance higher than in DAPHNIA treatment low ciliates rotifers suppressed enhanced bacteria abundance high ciliates high rotifers low bacteria abundance moderate ciliates moderate rotifers low bacteria abundance moderate ciliates moderate rotifers Daphnia Daphnia +F Filament Rotifer Copepod Figure 5a. F ina l zooplankton abundance in Summer enclosures. Values given are the mean of 3 replicates. Error bars indicate 1 standard error. Summer: Final Zooplankton Abundance in Enclosures 01 P a a 700 600 500 -400 300 --Daphnia Bosmina Copepod Treatment Rotifer 9 Bosmina I Daphnia QD Copepods 11 Rotifers I other Figure 5b. Fina l zooplankton abundance in the Fa l l enclosures. Values given are the mean of 3 replicates. Fall: Final Zooplankton Abundances in Enclosures 49 Figure 6a. Zooplankton biomass on the final day of the Summer experiment. Daphnia, Bosmina and other zooplankton biomasses were calculated using length-weight regressions (Table 5); rotifer biomasses were estimated using literature values (Table 6). Error bars indicate 1 standard error. Summer: Final Zooplankton Biomass • Bosmina 0 Daphnia • Copepods El rotifers • other zooplankton Daphnia Bosmina Copepod Treatment Rotifer Figure 6b. Zooplankton biomass on the final day of the Fa l l experiment. Daphnia, Bosmina and other zooplankton biomasses were calculated using length-weight regressions (Table 5); rotifer biomasses were estimated using literature values (Table 6). Fall: Final Zooplankton Biomass 900 800 700 „ 600 C/J /-—N | a 500 CQ <u S £400 3 1> 300 •-200 100 0 0 Daphnia Mcopepods H rotifers • other Daphnia Filaments Copepod Treatment Daphnia+F Rotifer 50 Summer Daphnia density was 89 + 28 individuals L"1 while the final Fall density was 74 + 5 individuals L" 1 in the Fall DAPHNIA and 35 + 11 individuals L"1 in the DAPHNIA+F treatment. High levels of reproduction of Daphnia resulted in a threefold increase in biomass over the course of the Summer experiment. Daphnia abundance in the Fall DAPHNIA treatment (74 + 5 individuals L"1 ) was similar to that observed in the Summer DAPHNIA treatment. The source populations in Summer and Fall differed in their size distributions (refer to Table 4 in Methods), and the number of Daphnia stocked per enclosure was increased in Fall to maintain similar Daphnia biomasses between seasons. This attempt to equalize the biomass could not influence subsequent reproductive behaviour of the Daphnia in response to seasonal differences in experimental conditions. Fortunately for the purposes of comparison, Daphnia pulex's population dynamics resulted in final population densities and biomasses that were nearly equal in the Summer and Fall DAPHNIA treatments (Figure 7). Despite the significantly lower mean weight of individuals in the final Daphnia population of the Fall DAPHNIA treatment (8.lug vs. 9.5 ug, A N O V A of DAPHNIA and DAPHNIA+F treatments: F ( 2 207g)=3.244, p=0.04, Tukey HSD comparison mean difference 1.3638, p= 0.04), the biomass of Daphnia pulex was not significantly different between any of the 3 treatments in which Daphnia were added (ANOVA F ( 2 6 )=2.465, p=0.16). The size and weight distributions for the Summer DAPHNIA and Fall DAPHNIA are given in Figures 8 and 9. None of the Daphnia in the Fall treatments attained the large sizes common in the Summer populations (>2.5mm) but there was a distinct cohort of adult animals larger than 1.5 mm in length, survivors from the initial zooplankton addition. Given that the mean length of Daphnia added to Fall enclosures was 1.5 mm at the 51 Figure 7. Abundance and biomass of Daphnia pulex in the 3 treatments in both experiments to which Daphnia were added. Error bars indicate 1 standard error. Daphnia Abundance and Biomass 1200 3 < SO 3 6 o £ 1000 800 600 --400 200 Summer Daphnia Fall Daphnia Treatment Daphnia + F • Number per litre • Biomass per Litre 52 Figure 8. Summer Daphnia length and weight distributions for individuals counted (samples pooled for allthree enclosures sampled. Summer Daphnia length distribution (0 •g > CD n E 175 150 125 100 %> %> W ^ _%> ^ Std. Dev = 439.31 Mean = 1064.0 N = 760.00 Length (um) 'o0 »oQ -o0 -oQ "oQ »o0 -oQ •o o o o o o o o o Summer Daphnia Weight distribution (ug) 300 Std. Dev = 12.31 Mean = 9.5 N = 760.00 °0 S0 'O <T i? ^ % • *0 7& && G0 <%• -fc ° ° o o o o o o o o o o o o o Weight (ug) N Std. Valid Missing Mean Deviation Statistic Statistic Statistic Std. Error Statistic LENGTH 760 0 1064.0112 15.9355 439.3113 Weight (ug) 760 0 9.4897 .4464 12.3056 53 Figure 9. Length and weight distributions of Daphnia in the Fall Daphnia samples (all three enclosures pooled).... Fall Daphnia length distribution 175 150 125 100 Std. Dev = 420.03 Mean = 994.4 N = 716.00 % ? ° o \ % 0 \ Length (um) •o •o o Fall Daphnia weight distribution 250 225 200 175 150 125 100 75 50 Std. Dev = 9.17 Mean = 8.1 N =716.00 ^mmmm^ w e i g h t y Std. N Minimum Maximum Mean Deviation Weight (ug) 716 .41 43.91 8.1175 9.1742 LENGTH 716 321.33 2254.57 994.4382 420.0318 Valid N (listwise) 716 54 start of the experiment, it is obvious that some growth of individuals had taken place (Figure 9). In both Summer and Fall experiments numerous juveniles were present. Despite the differences in the size and weight distributions of the Summer and Fall DAPHNIA treatments, they were equivalent in terms of biomass and from this perspective can be considered seasonal replicates. The abundance of Daphnia in the DAPHNIA+F treatment was lower and somewhat more variable between enclosures ( 35 ± 1 1 individuals L"1 ). Two enclosures in the DAPHNIA+F treatment had fewer than 25 Daphnia L"1 while a third had a Daphnia density similar to the Fall DAPHNIA enclosures. The difference in Daphnia biomass between the Fall DAPHNIA and DAPHNIA+F treatments was not significant (ANOVA of all DAPHNIA treatments: F ( 2 6 )=2.465, p= 0.16). Daphnia abundance was also lower in the DAPHNIA+F experiment than in the Fall DAPHNIA treatment, but not significantly so (ANOVA of all DAPHNIA treatments: F ( 2 6 )=6.601, p=0.03; Tukey HSD comparison for Fall DAPHNIA and DAPHNIA+F: mean difference = -38.5 individuals L"', p=0.1). This indicates that the presence of glass fibre filaments curtailed the growth of the Daphnia populations in the DAPHNIA+F enclosures, and the surviving individuals were larger in size than those in the Fall DAPHNIA (filament-free) enclosures (Figures 9 and 12). This difference in the size distribution is resulted in a difference in the mean weight of individual Daphnia between the two DAPHNIA treatments in Fall; 8.1 + .4 pg in the Fall DAPHNIA treatment and 9.1 + .4 pg in DAPHNIA+F. Biomass alone does not necessarily indicate equivalency for the Summer and Fall DAPHNIA treatments. Community filtering rate must also be considered. I used three regressions from the literature to generate estimates of the filtering capacity of Daphnia pulex 55 Figure 10. Estimated filtering capacities of Daphnia populations in experimental enclosures, expressed as the time required for the zooplankton to clear the entire enclosure volume of bacteria and algae. Fil tering capacity is estimated from the equations o f Haney 1985 and Petersen et al . 1978. Error bars represent 1 standard error. Comparisons of Population Filtering Capacities in all Daphnia Treatments b S "u C V s-£ o o s s a 10.00 9.00 8.00 7.00 6.00 5.00 4.00 3.00 2.00 1.00 0.00 • Summer Daphnia Fall Daphnia Treatment Fall Daphnia+F • Bacteria filter time (d) El Algae filter time (d) 56 in the experimental enclosures in all Daphnia treatments. The equations used are given in Table 7. The results are compared for all Daphnia pulex treatments in Figure 10, stated as the estimated time required by the Daphnia populations to clear all the water of algae or bacteria (bag turnover time). The final estimates of bag turnover time were log-transformed and compared using A N O V A , but no significant differences were detected, although in the case of filtering capacity on bacteria, the result approached significance (F ( 2 6 )= 4.418, p=0.07). However the observed power of the A N O V A was low (.526), mostly likely due to the large variance in estimated filtering capacities of the DAPHNIA+F enclosures. As Figure 10 illustrates, estimated enclosure clearance times are not appreciably different between the Summer and Fall DAPHNIA treatments. This indicates that the differences in the size distributions of the Daphnia pulex populations in the Summer and Fall DAPHNIA treatments do not translate into predictable differences in the potential grazing impact of Daphnia in the enclosures. The Summer and Fall treatments appear to be equivalent in their potential to influence the microbial and algal food webs if biomass and size distribution are used to predict their impact. 3.2 Daphnia and Filaments Initially the DAPHNIA+F treatment received the same stock population of Daphnia pulex as the enclosures in the Fall DAPHNIA treatment. Therefore, differences in the Daphnia population between the DAPHNIA+F and the Summer/Fall DAPHNIA treatments are the result of glass fibre filament addition and as such constitute a measurable treatment effect. The mean Daphnia biomass in this treatment, though lower than the other Daphnia treatments, was not 57 significantly different (see above) but the mean abundance of Daphnia in the DAPHNIA+F enclosures was significantly different from both the Summer and Fall DAPHNIA treatments (ANOVA: F ( 2 6 ) = 6.601, p=0.03; Tukey HSD test significantly different for the Summer DAPHNIA-DAPHNIA+F comparison, mean difference 53.67, p= 0.03). The population filtering capacity for algae and bacteria in this treatment was much lower (enclosure turnover time greater- Figure 10) and approached significance (see above). In two of the three DAPHNIA+F enclosures, Daphnia density was less than 25 individuals L" 1. There was very little increase in Daphnia abundance in the DAPHNIA+F treatment relative to the Fall DAPHNIA treatment. Though the mean length and weight per individual was significantly different between the Summer and Fall DAPHNIA treatments, this was notthe case in the DAPHNIA+F treatment (Summer DAPHNIA-DAPHNIA+F Tukey HSD comparison, mean difference in weight 0.35pg, mean difference in length, 23pm, p>0.58 for both). Figure 11 displays the weight and length distributions for the final DAPHNIA+F D. pulex populations. When compared to the Fall DAPHNIA treatment (Figure 9), the size distribution is skewed towards the larger size classes. This indicates that the presence of glass fibre filaments reduced the survival of Daphnia juveniles and/or reduced the reproductive rate of the adults. 3.3 Bosmina Final Bosmina abundances exceeded the target population density of 100 individuals L 1 and increased above the stocking density of 120 individuals L" 1. The final mean Bosmina density in enclosures was 188 + 17 individuals L"1 (Figure 5a). In comparison with the Summer DAPHNIA treatment, the population increase in BOSMINA over the course of the 58 Figure 11. Length and weight of Daphnia counted in final zooplankton sampled from the three Daphnia+F enclosures (pooled). Daphnia+F: Length distribution of individuals 175 T 1 150 125 100 Std. Dev = 442.07 Mean = 1040.7 N = 606.00 o Length (um) Daphnia+F: Weight distribution of individuals 2501 225' 200' 175-150-125-100 Std. Dev = 9.76 Mean = 9.1 N = 606.00 S> *> <$? <%> s= •<? -«P Weight (ug) N Minimum Maximum Mean Std. Deviation LENGTH Weight (ug) Valid N (listwise) 606 606 606 307.92 .37 2196.30 41.23 1040.7277 9.1392 442.0679 9.7606 59 Figure 12. Estimated filtering capacities of Bosmina and Daphnia populations in experimental, enclosures expressed as the time required for the zooplankton to clear the entire enclosure volume of bacteria and algae. Fil tering capacity is estimated from the equations of Haney 1985, Petersen et a l . 1978, and DeMot t 1982. Error bars represent 1 standard error. Comparison of Estimated Population Filtering Capacities in the Summer Daphnia and Bosmina treatments 20.00 19.00 18.00 -17.00 -16.00 15.00 -• 14.00 -13.00 12.00 11.00 10.00 9.00 8.00 7.00 - -6.00 5.00 - • 4.00 3.00 - -2.00 --1.00 -0.00 M-iM • Bacteria filter time (d) DI Algae filter time (d) Bosmina Summer Daphnia Treatment 60 experiment was modest. In terms of numbers this increase amounted to only 1.5 times the original stocking density, and therefore the final biomass observed is quite low relative to the zooplankton biomass in the DAPHNIA treatment. The estimated mean filtering capacity of the BOSMINA enclosures, in comparison with the DAPHNIA treatment, is shown in Figure 12. While the estimated turnover time for bacteria is quite long in the BOSMINA treatment, the estimated turnover time for algae is similar for the two treatments. This indicates that the BOSMINA treatment was able to "sweep" approximately the same daily volume of enclosure water as the Summer DAPHNIA treatment, but the two treatments differed in clearances rates for the algal and bacteria size fractions. 3.4 Copepods The final densities of adult copepods in the enclosures which originally received S. oregonensis additions was quite low (< 4 individuals L"1 ). The absence of adult copepods had been noted early in the experiment after visual assessment of enclosures during sampling. It is therefore assumed that S. oregonensis did not survive the handling procedure during the experimental set-up. For this reason, all 3 "copepod" enclosures in the Summer experiment were excluded from further analysis. This did not result in the total exclusion of a "copepod" grazer type from the Summer experimental design, however. While the adult copepods deliberately added to enclosures died as a consequence of handling, the nauplii inadvertently added to ROTIFER enclosures survived and reached maturity during the experiment. 61 Figure 13a. L i fe stage distribution of copepods in enclosures at the end of the Summer experiment. Values given are the mean of 3 replicate enclosures (2 in the Bosmina treatment). Error bars indicate 1 standard error. 250.00 „ 200.00 & 150.00 <D § 100.00 + 50.00 0.00 Summer: Final D. oregonensis abundance in enclosures -+-Daphnia Bosmina Copepod Treatment Rotifer • adult 3 copepodite E£3 nauplius Figure 13b. L i fe stage distribution of copepods in enclosures at the end of the Fa l l experiment. Values given are the mean of 3 replicate enclosures. Error bars indicate 1 standard error. 30 25 £ 20 a. 615 6 3 110 5 --Fall: Final D. oregonensis abundance in enclosures I adult I nauplius Daphnia Filament Copepod Treatment Daphnia +F Rotifer 62 The copepod densities in the "ROTIFER" enclosures are given in Figure 13a-b. Two of the Summer ROTIFER enclosures had copepod numbers in excess of 150 individuals L"'. The mean copepod density for the entire treatment was 156.9 ± 110.2 individuals L" 1. There were also copepods present in the Fall ROTIFER enclosures (Figure 13b), though at much lower abundance than in Summer. The majority of copepods present in the ROTIFER enclosures were adult or late stage copepodites; animals of this size would have been excluded from the enclosures by the initial filtration of enclosure water. Therefore, adult copepods present in these enclosures at the end of the experiment must have been added as nauplii contaminating the initial "rotifer" stock. The high copepod density in the Summer "ROTIFER" treatment had changed the nature of the grazer community from rotifer-dominated to copepod-dominated . In order to reflect this change, the former "ROTIFER" treatments have been renamed as "COPEPOD" treatments in what follows. The Summer COPEPOD treatment had high copepod abundances, but the Fall COPEPOD treatment had very low copepod abundances, so that in effect the "Fall COPEPOD" treatment is very similar (in terms of the grazer community) to the "Fall ROTIFER" treatment discussed below. The estimated filtering capacities (on flagellates and ciliates) for the Summer and Fall COPEPOD treatments are given in Figure 14. These estimates are based on per capita clearance rates measured for S. oregonensis (Sanders and Wickham 1993). S. oregonensis is not bactivorous, however its clearance rates on ciliates can be high. The copepod population would have been able to "sweep" the same enclosure volume per day as the DAPHNIA and BOSMINA treatments, if grazing on ciliates. This is not the case for the Fall COPEPOD treatment, where copepod density was low. Therefore the estimated enclosure "turnover" 63 Figure 14. Estimated time required by the copepod population to filter all of the enclosure volume, using flagellates and ciliates as reference prey items. Error bars indicate 1 standard error. Comparison of copepod filtering capacities on protista H Flagellate turnover time (d) • Ciliate turnover time (d) Summer Copepod Fall Copepod Treatment 64 time for this treatment was quite long and not comparable to the DAPHNIA and BOSMINA treatments. 3.5 Rotifers At the end of the Summer experiment, I became concerned about my ability to evaluate the successful implementation of the "rotifer" treatment. I had therefore included an informal comparison "treatment" in the Fall experiment which was designed to provide information on the success of rotifer population additions. Two additional enclosures received heat-killed rotifer inoculum concurrent with the addition of live inoculum to experimental enclosures. At the midpoint of the Fall experiment, the mean rotifer abundance in the killed controls was not significantly different from enclosures where live rotifers had been added (t= 1.456, df=3, p=0.24, see Figure 15a). The killed controls were also sampled for rotifers on the final day of the experiment. Rotifer densities in enclosures with no zooplankton added were not different from either the enclosures with rotifers added or the "killed rotifer" controls (Figure 15b). The rotifer abundances at the midpoint of the fall experiment apparently reached the target density of 1000 ind L"1 before declining to the levels observed in the final samples, but this was not related to the addition of rotifers to the enclosure bags. The results above suggest that the addition of live rotifers to enclosures was not likely responsible for observed rotifer densities at the end of the Fall experiment. The similarity of the treated enclosures to the "no grazer added" and "killed control" enclosures at the midpoint of the Fall experiment suggests that the added rotifers died soon after inoculation. As the 65 Figure 15a. Rotifers at the midpoint in the Fa l l experiment. The animals were added to the Rotifer enclosures (3 replicates) at densities of 1200 individuals per litre; while the K i l l e d Rotifer enclosures (2 replicates) recieved heat-killed inoculum. Error bars indicate one standard error. Midpoint Estimate of Rotifer density in Copepod (rotifers added) and Killed Rotifer (heat-killed rotifers added) Treatments 2000 1800 -1600 „ 1400 + u 3 1200 u g 1000 S 800 o at, 600 400 -200 -0 Copepod Killed Rotifers Treatment Figure 15b. Rotifer densities in the Fa l l experiment. The animals were added to the Rotifer enclosures at densities of 1200 individuals per litre; rotifers were init ially reduced by filtering, while the K i l l e d Rotifer enclosures (2 replicates) received heat-killed rotifer inoculum. Error bars indicate 1 standard error. A comparison of final rotifer densities in treatments with rotifers added or excluded (Fall treatments) 600.00 500.00 | 400.00 + u g 300.00 1 200.00 100.00 --0.00 Copepod Rotifer Treatment Killed Rotifers 66 Summer "no grazer" treatment had the same mean rotifer abundance as treatments where live rotifers were added, it is likely that the rotifer additions in the first experiment also had no effect on final rotifer abundance. The total rotifer abundance in the Summer ROTIFER enclosures was 170 + 33 individuals L" 1, substantially below the 1000 L~' stocking density. Furthermore, rotifers were present in all enclosures except those which contained Daphnia. Initial rotifer populations were drastically reduced, but not eliminated, by filtration. By the end of both Summer and Fall experiments, the rotifer populations had increased. These rotifer populations somewhat confound all the grazer treatments, but their biomass and filtration capacity (not shown)4 was likely much lower than the DAPHNIA, BOSMINA and COPEPOD grazer populations. Only those enclosures which received no zooplankton additions could be considered "rotifer-dominated" at the end of the experiment (but the copepod presence in the Fall COPEPOD treatment is very limited and this community is effectively rotifer-dominated as well). In the Summer and Fall "NO GRAZER" treatments, rotifers constituted the dominant fraction of the metazoan grazer community (Figure 6a and 6b). These treatments are therefore referred to as "ROTIFER" treatments in the text below. This signifies that there is no treatment in the design which completely excludes all but the microbial grazers. The estimated filtering capacities of K. cochlearis and P. vulgaris on nanoplankton and bacteria vary by an order of magnitude in the literature (Sanders et al. 1994). Some estimates of P. vulgaris'?, clearance rate on bacteria would indicate potential enclosure Filtration capacity was estimated using per capita rates available in the literature, but the range of possible values (for both K. cochlearis and P. vulgaris) was quite large, and I chose not to present the results in detail. Clearance rates for rotifers can range from 0.001 to 0.072 ml rotifer"' hr"1 (Sanders et al . 1994). 67 turnover times of more than 130 years for some rotifer populations in the Fall experiment. However, a measured clearance rate found in another study for a mixed Keratella'Polyarthra community in situ (Sanders, et al. 1994) would give respectable turnover times on the order of 3-5 days for the rotifer populations in the Summer Enclosures, when feeding on nanoflagellates. Comparable values were obtained for the Fall experiment. Based on the wide range of literature values, the equivalency of "enclosure turnover time" between the ROTIFER enclosures and the other treatments cannot be reliably assessed without measured grazing rates. It is within the realm of possibility that the Summer and Fall ROTIFER treatments were characterized by grazing pressures on nanoplankton roughly equivalent to those in the macrozooplankton-dominated enclosures. However, the enclosures containing Bosmina and S. oregonensis populations have community clearance rates which combine those of macrozooplankton and rotifers, and are therefore higher than those in the Summer and Fall ROTIFER treatments. 3.6 Other zooplankton Figure 5a indicates that in all treatments there are a few zooplankters characterized as "other" (their biomass is also indicated in Figure 6a). A few individuals of Sida crystallina (O.F.Miiller) 1875, Diaphanosoma brachyurum (Lieven) 1848, Chydorus sphaericus (O.F.Miiller), Simocephalous vetulus Sch0dler 1858 , Ceriodaphnia sp. Dana 1853 and the occasional ostracod were present in some final zooplankton samples. Their appearance in enclosures was sporadic and unrelated to the treatments; it is likely that such individuals escaped into the enclosures during the filtering process (perhaps as eggs), or perhaps were 68 Figure 16a. N o o n temperatures measured in the Summer experiment (prior to daily sampling). Summer Temperature in Pond 13 enclosures 0 "I 1 1 1 1 1 1 ; 1 1 1 1 1 1 1 1 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 Julian Day Figure 16b. N o o n temperatures measured in the F a l l experiment (prior to daily sampling). Fall Temperatures in Pond 13 69 added accidently along with the treatment zooplankton. As these uninvited guests were never abundant, they were not important components of the grazer community in any of the experimental enclosures. Still fewer of these stray zooplankton were present in the Fall enclosures. 3.7 Temperature Daily temperature measurements were part of the sampling protocol in both Summer and Fall experiments, however due to a malfunction of the probe used to determine pH and temperature, some measurements are missing from the Summer experiment. The results obtained are presented in Figure 16a-b. Weather conditions were generally cloudy with light showers for the first half of the Summer experiment, and water temperature was stable near 19°C. The final week of the experiment was characterized by hot, sunny weather and resulted in a warming trend with enclosure temperatures rising to 21.5°C. The temperature pattern in the Fall experiment was much different. The Fall experiment was set up during a period of warm weather and water temperature was 15°C. Coincident with the onset of sampling, the weather became much colder with frequent episodes of rain. Water temperatures declined gradually throughout the experiment, to a low of6°C. The pH measurements in enclosures in both experiments ranged between 8 and 8.9, and on most dates were approximately 8.6, with very small variability between enclosures on any given day. 70 4. Part II : Numerical responses to grazer manipulations 4.1 Bacterial abundance To place the observed variation in bacterial abundance in perspective, Figure 17 shows the range of bacterial abundance (all individual estimates pooled) for both Summer and Fall experiments. The minimum and maximum values are given in Table 9. The range observed experimentally is compared to literature values of bacterial abundance recorded from a variety of freshwater ecosystems. The range of variation in bacterial abundance observed in this study encompasses a large fraction of the variation in bacteria density across diverse freshwater ecosystems. The literature values chosen consist mainly of seasonal minima and maxima for particular lakes, though some observations may also have been taken over shorter periods of time. Year to year variation within lakes is also included in the literature data set. The lowest single estimate of bacterial abundance observed in my experiments was 9.1 x 105 cells ml"1 in the Fall experiment, while the highest observed was 6.8 x 106 cells ml"1 in the Summer. A sample taken from open water at the centre of Pond 13 on August 12 (at the same depth as the enclosures) was 2.29 x 106 cells ml"1. The enclosures appear to be appropriate models of the natural bacterial abundance in Pond 13. 4.1-1 Bacterial abundance- Summer Bacterial abundances observed in replicate enclosures of each treatment in the Summer experiment are shown in Figure 18a, b, c and d. The sample dates included in the statistical analysis are indicated in Table 8a. The results of the A N O V A procedure are summarized in 71 Figure 17. Bacterial abundance measurements Boxplots of experiments and literature values Estimates from each experiment are pooled Summer experiment 75 Fall experiment 46 Literature values Source of abundance measurements Literature: pooled high and low values from freshwater habitats Blue line represents mean; error bars indicate range of values References: Bennet et al. 1990, Berninger et al. 1991Bird and Kalff 1984, Gude 1988, Gude 1991, Hardy etal. 1986, Markosova and Jezek 1993, PsennerandSommaruga 1992, Weisse 1990. 72 Table 9. Range of bacterial abundance estimates observed during the two experiments in this study. A comparison with a range literature values is shown in Figure 18. Values are given as cells ml"1. Experiment # of minimum maximum mean Std. Error observations Summer 54 1.6 X 106 6.8 X 106 3.6 x 106 1.6 x 105 Fall 75 9.1 X 105 4.3 X 106 2.4 x 106 9 x 104 Both seasons 129 9.1 X 105 6.8 X 106 2.9 x 106 1 x 105 73 Figure 18. Bacterial cell numbers in Summer enclosures are shown. Points represent individual enclosures. The Dark line represents the mean of 3 replicates for the Daphnia, Copepod and Rotifer treatments. The Bosmina treatment has 2 replicates. 74 Appendix 1. The within-subjects effects indicate a significant effect of sample date (F { 4 2 8 )= 12.54, p<.001), and a significant interaction of date and treatment (F ( 1 2 2 8 ) = 6.472, p<.001). The between-subjects effects indicate a significant effect of treatment (F ( 3 7 )= 32.586, p<.001). Error variances were unequal for two dates at the end of the experiment (Appendix 1, Table C), and a in this case may differ from the stated level of 0.05. However, differences between treatments are obvious from visual inspection of Figure 18. As an example, mean bacterial abundance between the DAPHNIA and COPEPOD treatments differs by 2.5 times on the final day of the experiment and the abundance values for individual enclosures do not overlap. I consider this difference to be real and biologically significant. Differences between particular treatment means were examined post-hoc using Tukey HSD comparisons. A l l treatments were compared on each sample date; the results obtained are summarized in Table 10. No treatment differences were discernable prior to August 17 (Julian day 229, the midpoint of the experiment), and so the table displays only the results for sample dates where significant differences between treatments were detected. The four treatment means are shown together in Figure 19 for all dates included in the repeated measures A N O V A . It is apparent that bacterial abundance in the COPEPOD treatment remained at or near the same level for the duration of the experiment. Bacterial abundance in the DAPHNIA and ROTIFER treatments declined throughout the experiment and was significantly less than the COPEPOD and BOSMINA treatments when the experiment was concluded. The BOSMINA treatment showed an initial decline in abundance similar to that of the DAPHNIA treatment, but after August 17 bacterial abundance in this treatment increased to match that observed in the COPEPOD treatment. 75 Table 10. Tukey multiple comparison test results fo l lowing repeated measures A N O V A of natural log-transformed bacteria abundances (5 dates) in the Summer experiment. N o comparisons before August 17 0 1 (August 9, August 12) were significant (not shown). Significant differences are indicated in bold type, oc= 0.05 Date Treatment Treatment Mean difference Standard S ig . (i) 0) ( i - j ) Error August 17 Daphnia Bosmina .1957 .125 .452 Copepod -.3769 .112 .046 Rotifer .4765 .112 .015 Bosmina Copepod -.5726 .125 .011 Rotifer .2808 .125 .200 Copepod Rotifer .8534 .112 .001 August 23 Daphnia Bosmina -.7467 .220 .045 Copepod -.8021 .197 .028 Rotifer .01524 .197 1 Bosmina Copepod -.0554 .220 .994 Rotifer .7619 .220 .041 Copepod Rotifer .8173 .197 .018 August 24 Daphnia Bosmina -.7420 .119 .002 Copepod -.9230 .106 <.001 Rotifer -.2843 .106 .115 Bosmina Copepod - -.1811 .119 .473 Rotifer .4577 .119 .025 Copepod Rotifer .6387 .106 .002 76 Figure 19. Mean of natural log-transformed bacteria abundance for the Daphnia, Bosmina, Rotifer, and Copepod treatments in the Summer experiment. Homogeneous subsets are indicated by ellipses. Julian Day 77 4.1-2 Bacterial abundance- Fall The repeated measures A N O V A of natural log-transformed bacterial abundance in the Fall experiment includes 5 sample dates (Table 8b). The A N O V A results are given in Appendix 1 Table D, E, and F. The within subject effect of sample date (F ( 4 4 0 ) = 13.474, p<0.001) and the date x treatment interaction (F ( 1 6 4 0 ) = 5.348, p<0.001) are significant. There is also a significant effect of treatment (F ( 4 1 0 ) = 18.426, p<0.001). Again, despite log-transformation of the data, variances were not equal for all sample dates (Appendix , Table F). However, this assumption of the A N O V A is only violated for a single sample date and A N O V A is generally thought to be robust to violations of this assumption in many circumstances (for a discussion see Winer et al. 1991 or Kirk 1982). The bacterial abundances for each Fall treatment are shown in Figure 20a-e. The effect of Daphnia on bacterial abundance in this experiment is immediately apparent. Bacterial abundance in the DAPHNIA treatment remained constant and a modest increase occurred in the DAPHNIA+F treatment. This is in stark contrast to the COPEPOD, ROTIFER and FILAMENT treatments, which show modest increases in bacterial abundance up to the mid-point of the experiment, followed in each case by a steep decline. This pattern also contrasts markedly with that observed in the Summer experiment, in which the COPEPOD treatment had a constant (and high) bacterial abundance while the DAPHNIA treatment showed a decline. Tukey HSD post-hoc comparisons of individual treatments (on each sample date) did not detect any differences among treatments prior to November 2 (Julian day 306); this is also apparent in visual inspection of Figure 20. The results obtained from the Tukey HSD multiple comparisons are given in Table 11 for the dates where significant effects were 78 Figure 20. Natural log-transformed bacterial cel l numbers in F a l l enclosures are shown. Points represent individual enclosures. The dark line represents the mean of 3 replicates for the Daphnia, Daphnia+F, Copepod and Rotifer treatments. 20a. Bacteria Abundance in Fall Daphnia Enclosures 15.5 15.3 -15.1 -14.9 14.7 14.5 14.3 14.1 13.9 13.7 •• 13.5 - 4 - - 4 - - 4 - - 4 - -+-290 292 294 296 298 300 302 304 306 308 Julian Day 20b. Bacteria Abundance in Daphnia+F Enclosures 15.5 15.3 15.1 14.9 14.7 --14.5 14.3 14.1 + 13.9 13.7 13.5 - 4 - - 4 - -+- -+- -+• - 4 -290 292 294 296 298 300 302 304 306 308 Julian Day 20c. c -I Bacteria Abundance in Fall Copepod Enclosures 15.5 -| 15.3 • 15.1 - A 14.9 -- • A 14.7 - • s 14.5 • A 14.3 - \ 14.1 - •\ 13.9 - \ 13.7 - • 13.5 - 1 h 1 1 h 1 h -— h — — 4 — 290 292 294 296 298 300 302 304 306 308 Julian Day 20d. Bacteria Abundance in Fall Rotifer Enclosures 290 292 294 296 298 300 302 304 306 308 Julian Day 20e. Bacteria Abundance in Filament Enclosures s 290 292 294 296 298 300 302 304 306 308 Julian Day 79 Table 11. Tukey multiple comparison test results based on repeated measures ANOVA of natural log-transformed bacterial abundances (5 dates) in the Fall experiment. No comparisons before November 2 n d (October 18, October 25, October 30) were significant (not shown). Significant differences at a= .05 are indicated in bold type. Date Treatment Treatment (j) Mean difference Standard Sig. (i) 0 - j) Error November 2 Daphnia Filament .2901 .154 .385 Copepod .3900 .160 Daphnia+F -.2456 .533 Rotifer .2926 .378 Filament Copepod .0999 .963 Daphnia+F -.5357 .038 Rotifer .00246 1 Copepod Daphnia+F -.6356 .014 Rotifer -.0974 .966 Daphnia+F Rotifer -.5382 .037 November 4 Daphnia Filament .5937 .100 .001 Copepod .9240 <.001 Daphnia+F -.2778 .112 Rotifer .8358 <.001 Filament Copepod .3303 .050 Daphnia+F -.8715 <.001 Rotifer .2421 .189 Copepod Daphnia+F -1.2018 <.001 Rotifer -.0882 .899 Daphnia+F Rotifer 1.1136 <.001 80 detected. The FILAMENT, ROTIFER and COPEPOD treatments are significantly different from DAPHNIA+F on November 2 n d; the DAPHNIA treatment is significantly different from DAPHNIA+F at an a of 0.053. By the final day of the experiment, the DAPHNIA and the DAPHNIA+F treatments are significantly different from the other treatments but not from each other (Figure 21). Additionally, the FILAMENT treatment has a final bacterial abundance significantly higher than that observed in the COPEPOD and ROTIFER treatments. 4.1-3 Bacterial abundance- Seasonal comparison of DAPHNIA, COPEPOD and ROTIFER The three treatments performed in both experiments can be compared to illustrate the seasonal differences in their grazing impact on the microbial web, as implied by changes in bacterial abundance. Figures 22, 23 and 24 depict bacterial abundance under the repeated treatments in both experiments. The data given are natural log-transformed and each point represents the mean of three replicates. The Summer experiment took place over 16 sampling days, the Fall experiment over 18; for ease of comparison abundances are given according to the time elapsed in each experiment rather than by Julian date. Repeated measures A N O V A could not be performed on these data, because too few samples were taken at the same time relative to the onset of sampling in each experiment. Bacterial abundance as a whole was lower in the Fall than in the Summer (t-test on log-transformed abundance measurements pooled for each season, t=6.662, df=157, oc=.05, p<.001). There were large differences in water temperature between the Summer and Fall experiments. Despite this, when the time course of abundance changes are compared, significant differences among treatments in either experiment are only apparent after the midpoint. 81 Figure 21. Mean of natural log-transformed bacteria abundance for the Daphnia, Daphnia+F, Filament, Copepod and Rotifer treatments in the Fall experiment. Homogeneous subsets as determined by Tukey multiple comparisons are given by ellipses. See Appendix 1 Table G for significance levels. 155, , 290 295 3D0 35 3D Julian Day 82 Figure 22. A comparison of natural log-transformed bacteria abundances in the Summer and Fall Daphnia treatments. The Summer experiment lasted 1 6 days, the Fall experiment 18. Data points represent the mean of three replicates. S-H CD & , 03 •c CD • 4 — * o a 16 15.5 15 14.5 14 13.5 f Summer Daphnia • Fall Daphnia 10 12 14 16 18 20 Day of Experiment Figure 23. A comparison of natural log-transformed bacteria abundances in the Summer and Fall Rotifer treatments. The Summer experiment lasted 16 days, the Fall experiment 18. Data points represent the mean of three replicates. •c • 4 — * o H-l 16 15.5 15 14.5 14 --13.5 Summer Rotifer Fall Rotifer 10 15 20 Day of Experiment 83 Figure 24. A comparison of natural log-transformed bacteria abundances in the Summer and Fall Copepod treatments. The Summer experiment lasted 16 days, the Fall experiment 18. Data points represent the mean of three replicates. 16 135 ^ 1 1 1 1 0 5 10 15 2D Day of Experiment 84 Figure 22 illustrates the decline in bacterial abundance observed in the Summer DAPHNIA experiment relative to the gradual increase seen in the Fall. While overall bacterial abundance does not differ markedly between the two seasons, the maintenance of a high bacterial abundance in the DAPHNIA treatment in the Fall experiment occurs in opposition to the general seasonal trend to lower bacterial numbers. The bacterial abundance pattern in the ROTIFER treatment is shown in Figure 23. The Summer treatment shows a gradual decline in abundance, followed by a slight recovery towards the end of the experiment. The Fall pattern shows a gradual increase in abundance, followed by a sharp decline. This pattern in bacterial abundance is similar to that observed in the FILAMENT (Figure 20e) and COPEPOD (Figure 20c) treatments in the Fall experiment. There is a decline in bacteria standing stock observed for the ROTIFER treatment in both seasons. In contrast, the COPEPOD treatment exhibits high bacterial abundance in Summer, and a decline in bacterial abundance in Fall (Figure 24). The Summer COPEPOD enclosures were S. oregonensis-dominated, but copepod populations did not increase in Fall enclosures to the same extent as in Summer. The Fall COPEPOD and Fall ROTIFER enclosures thus had similar zooplankton communities and exhibited the same trend in bacterial abundance, while the Summer COPEPOD enclosures had bacterial abundances similar to the BOSMINA enclosures. 4.2 Response of rotifers to treatments In addition to the initial attempts to manipulate rotifer biomass in what eventually became the "COPEPOD" treatment, rotifer abundance and species composition in the DAPHNIA, 85 BOSMINA, ROTIFER and FILAMENT treatments exhibited a response to the various zooplankton additions (or lack thereof). Due to the mesh size used to filter enclosure water, low numbers of rotifers were initially present in all bags. Rotifer abundances in the enclosures on the final day of the Summer experiment are given in Figure 25a. Growth of the rotifer population was suppressed in the DAPHNIA treatment. The COPEPOD enclosures had rotifer populations similar to the ROTIFER enclosures, despite the addition of 1200 rotifers L"1 to the former. There appears to be a slight enhancement of rotifer numbers in the BOSMINA enclosures. The results for the Fall enclosures are similar, with the two Daphnia enclosures showing greatly reduced rotifer numbers, while the Fall COPEPOD treatment shows no increase in rotifer densities over that observed in the ROTIFER treatment. Table 12a gives the results of a one-way A N O V A comparison of log-transformed rotifer abundances across all treatments in the Summer and Fall experiments. There is a significant effect of treatment (F ( 8 1 7 )= 7.926, p< .001), however despite log-transformation of the data, the variances were not homogenous (Table 12b) and therefore the reported p-values may be inaccurate. Differences between specific treatments were determined in post-hoc testing using the Tukey HSD procedure; the significance levels of the testing outcomes are given in Appendix 1 Table G. Log-transformed abundances are displayed graphically in their homogeneous subsets in Figure 26. The comparison of rotifer abundances across all treatments in both seasons indicates a strong inhibition of rotifer population growth in the presence of Daphnia. This effect of Daphnia appears to be ameliorated by the presence of glass fibre filaments in the DAPHNIA+F 86 Table 12a. A N O V A comparison of natural log-transformed rotifer abundance on the final day of sampling. Includes all treatments from both the Summer and Fall experiments, °c= .05 Source Sum of df Mean Square F Sig. Squares Treatment 29.634 8 3.704 7.926 <.001 Error 7.945 17 .467 Total 37.579 25 T R squared = .789 (adjusted R squared = .689) Table 12b. Levene's test of equality of error variances: tests the null hypothesis that the error variance of the natural log-transformed rotifer abundance is equal across treatments. The dependent variable is the natural log-tranformed rotifer abundance on the final day of sampling. Treatments from both the Summer and Fall experiments were included. Variable F df 1 df 2 Sig. Rotifer Abundance 2.978 8 17 .028 87 Figure 25a. Total rotifer abundance on the final day of the Summer experiment (all species). Abundance values are given as a mean of 3 replicates (2 replicates in the Bosmina treatment). Rotifer density in Summer Enclosures 700 600 •• Daphnia Copepod Rotifer Bosmina Treatment Figure 25b. Total rotifer abundance on the final day of the Fa l l experiment (all species). Abundance values are given as a mean of 3 replicates per treatment. 88 Figure 26. Natural log-transformed rotifer abundance on the final day of sampling both experiments. Homogenous subsets (Tukey multiple comparison, Appendix 1 Table G) are indicated by the solid lines. Note that the Daphnia+F treatment is significantly different from the Summer Daphnia at the 0.053 level) 8 7 0 J 1 1 1 1 1 1 1 1 1 1 8 1 1 1 1 1 1 1 1 1 — — 1 1 1 1 1 1 - J — Surrrrer Fall C&phnia Surrmer Surrrrer Fall Fall filament Bosrrina Daphnia Daphnia +F Copepod Rotifer Rotifer Copepod Treatment 89 treatment (Fig. 27), which has rotifer abundances similar to the ROTIFER and COPEPOD treatments in Summer and Fall. Rotifer densities tended to be higher in the Fall treatments, but the highest abundances were recorded in the Summer in the Bosmina enclosures. While the patterns in rotifer abundance remained similar across treatments in both the Summer and Fall experiments, the species composition of the rotifer community differed between seasons. Figure 27 illustrates the major change in species composition. In Summer Keratella cochlearis was most abundant while in the fall Polyarthra c.f. vulgaris were more numerous. Lecane spp. (2 species) were also relatively abundant in the Summer COPEPOD enclosures at a mean density of 81 individuals L" 1. Lecane spp. were present in the other Summer enclosures at low densities (< 15 individuals L" 1 ), but were not recorded in the Fall zooplankton samples. Keratella quadrata was present in some enclosures at very low densities (~ 1 individual L"1 ), absent from others, and was never abundant. 4.3 Response of ciliates Ciliates were present in all enclosures, and there were significant differences in ciliate abundance across treatments in both experiments (ANOVA: F { 8 1 7 ) = 6.838, p<.001) . In a pattern similar to that observed in rotifer densities, ciliates abundance is lowest in the Summer and Fall DAPHNIA treatments (Figure 28a and 28b). The highest ciliate densities were recorded in the ROTIFER and FILAMENT treatments. Contrasts between all treatments were compared using the Tukey multiple comparison procedure, and the homogenous subsets are shown graphically in Figure 29. The Summer and Fall DAPHNIA, Summer COPEPOD, and BOSMINA treatments all had low final ciliate abundances. The lowest ciliate density was 90 Figure 27. Abundance estimates of the most common rotifer species in both the Summer and F a l l experiments. Values given are the mean of 3 replicates for each treatment (2 i n the Bosmina treatment). These two species comprised most o f the rotifer populations in enclosures. Comparison of Final Rotifer species composition in Enclosures: Summer and Fall 600 -i : 1 Daphnia Copepod Daphnia Copepod Rotifer Summer Summer Fall Fall Fall Treatment 91 Figure 28a. Cil iate abundance in the Daphnia, Bosmina, Rotifer and Copepod treatments on the final day of the Summer experiment. Densities given are means of 3 replicates (2 in the Bosmina treatment). Error bars indicate 1 standard error. Summer Final Ciliate Abundances in Enclosures 12000 b 10000 8000 a a. £ 6000 CO s 4000 c 5 2000 Daphnia Copepod Bosmina Treatment Rotifer Figure 28b. Cil iate abundance in the Daphnia, Daphnia+F, Filament, Rotifer and Copepod treatments on thefinal day of the Summer experiment. Densities given are means of 3 replicates. Error bars indicate 1 standard error. Fall Final Ciliate Abundances in Enclosures 12000 10000 --= 8000 a> a. 10 £ 6000 u I 4000 2000 Daphnia Daphnia+F Copepod Treatment Rotifer Filament 92 Figure 29. Natural log-transformed ciliate abundances on the final day of the S Summer and Fall experiments. Homogenous subsets (Tukey mutliple comparison) are shown by the black bars. 12 f 10 •-perli 8 --:es 6 -Lliat 4 --o 9 - -LN c. 0 --I n n — * j n -+-• i i BH I Summer Fall Summer Bosmina Daphnia+F Fall Fall Rotifer Summer Filament Daphnia Daphnia Copepod Copepod Rotifer Treatment 93 Figure 30a. Species composition of the ciliate community in the Summer enclosures. Densities are given as the mean of 3 replicate enclosures (2 i n B o s m i n a ) Identification is made to the Order level except where indicated. Summer Ciliate Abundance in Enclosures 6000 5000 'f 4000 -- I Strombidium sp. • Gymnostomatida • Hypotrichida 1 H Miscellaneous Daphnia Bosmina Copepod Rotifer Treatment Figure 30b. Species composition of the ciliate community in the Fa l l enclosures. Densities are given as the mean of 3 replicate enclosures. Identification is made to the Order level except where indicated. Fall Ciliate Abundance in Enclosures 6000 y 5000 - -I Strombidium sp. • Gymnostomatida • Hypotrichida 1 H Hypotrichida 2 fflDiscomorpha sp. B Miscellaneous Daphnia Filament Copepod Daphnia+F Rotifer Treatment 94 found in the Summer D A P H N I A treatment, which was significantly different from the D A P H N I A + F , Fall C O P E P O D , Fall R O T I F E R , Summer R O T I F E R and F I L A M E N T treatments. Ciliates were identified to the Order level in most instances. In the majority of the ciliate sub-samples, total ciliates counted numbered less than 200 (in a 25 ml settling chamber). Counts of this magnitude are acceptable for an assessment of total ciliate abundance. The counts of particular species are small subsets of the total count, and are too low in absolute numbers to allow accurate estimates of their density. With this caveat, the relative abundances of the most abundant ciliate groups identified are shown in Figure 30a (Summer) and Figure 30b (Fall). The most commonly observed species observed in both Summer and Fall experiments was Strombidium sp. Other ciliates observed were counted and described as "morphotypes" and later identified. The Gymnostomatida, Hypotrichida type 1 and Hypotrichida type 2 ciliates observed were all single species, and in the case of the Gymnostomatida and Hypotrichida 1, the same species was observed in both experiments. Hypotrichida 2 was also relatively common in the Fall experiment but was not observed in Summer. A species of Discomorpha sp. was also frequently observed in the Fall enclosures, but only in low relative abundance. In general ciliate abundance was higher at the end of the Fall treatments (excluding Daphnia treatments), and higher treatments where large metazoan grazers had been excluded. The Fall increase in abundance is largely due to an increase in the most common species, Strombidium sp. (Figures 30a and 30b). 95 4.4 Response of algae Lugol-preserved samples were settled for the counting of ciliates; algae were not enumerated. In the process of counting ciliates, however, some general observation of the algal abundance and diversity in enclosures were made. Lugol-preserved samples from the final sampling date of each experiment were examined. The flora present remained typical of that observed in Pond 13 in a pilot study conducted in May 1995. Algae samples from Pond 13 were visually inspected prior to the Summer experiment using an inverted microscope at 100X magnification and the composition remained typical of that observed in the late spring. The most common algae in Pond 13 are small chrysophytes and small cryptomonads. In the spring the green alga Selenastrum sp. was present in high abundances but was rarely observed in the samples from the Summer and Fall experiments. Another gelatinous green alga, Elakatothrix sp. was present in the spring and in the Summer experiment, though at relatively low abundance. Typically there are dinoflagellates such as Ceratium sp. and other large algal species present in Pond 13. These cells were excluded from the enclosures by the initial filtration and were present only in low numbers. In the case of Ceratium, cells were observed at high abundance in the Pond during initial surveys of microzooplankton prior to the experiment. Large Volvox colonies were also present in the Pond prior to the Summer experiment and were noted in the zooplankton tows taken from Pond 13 during the harvesting of Bosmina and S. oregonensis. Arthrodesmus sp. and Neurocytium sp. were present in the spring and in the Summer enclosures, but at very low abundance. 96 In the Summer enclosures, there was a particularly visible effect of the DAPHNIA treatment on the algae present. A large bloom of Volvox occurred in Summer in all the Daphnia enclosures (2940 + 355 colonies L" 1 ) which was not observed in the other treatments. This bloom turned the water in the enclosures a murky green while the other bags remained clear. Inspection of the samples indicated that there were few edible algal cells in the DAPHNIA enclosures on the last day of the experiment; only broken Volvox colonies and a few cells (and large ciliates) were present in the samples. In contrast, the samples from the other treatments contained abundant edible algae in Daphnid? preferred feeding range. Volvox blooms also occurred in the Fall DAPHNIA enclosures, but the intensity of the water colour never reached the deep green observed in Summer. In counts of colony densities on the final sampling day of the Fall experiment, Volvox densities were 465 + 35 colonies L" 1 in the DAPHNIA enclosures and 577 ± 125 colonies L"1 in the DAPHNIA+F enclosures. Volvox colonies were present in the other enclosures, but were not abundant. In contrast to the Summer DAPHNIA enclosures, Fall enclosures containing Daphnia also had small edible algal cells present. Abundance of edible cells appeared to be somewhat less than that observed in the other Fall enclosures, but did not approach the "clear water" state seen in the Summer DAPHNIA treatment. However, algal diversity was somewhat lower in the Fall enclosures due to the appearance, of an algal bloom described below. Inspection of the algal samples from the final day of the Fall experiment indicated that a bloom of Elakatothrix sp. was present in all the enclosures. This gelatinous green alga can form sheets, but was present in the samples as single cells, although occasionally two or more 97 Figure 31. C e l l numbers of Elakatothrix sp. in the Fa l l enclosures. The densities given for the Daphnia, Daphnia+F, Rotifer, Copepod and Filament treatments are mean values for 3 replicate enclosures. Error bars represent 1 standard error. Density of Elakatothrix sp. in Fall Enclosures 2.50E+04 2.00E+04 - • a 1.50E+04 ~ 1.00E+04 U 5.00E+03 - • 0.00E+00 Daphnia Daphnia +F Rotifer Treatment Copepod Filament 98 Table 13. A N O V A of natural log-transformed Elakatothirx sp. density on the final day of the Fall experiment. Source Sum of df Mean Square F Sig. Squares Treatment 2.305 4 .576 11.715 .001 Error .492 10 .04918 Corrected Total 2.796 14 R squared = .824 (Adjusted R squared = .754) 99 Figure 32. Natural log-transformed Elakatothrix densities in the Fal l experiment. Homgenous subsets according to the Tukey multiple comparison procedure are indicated. The Daphnia -Rotifer comparison is significantly different at 0.053, all other comparisons are significant at <.05. 11 1 ^10 Daphnia Daphnia+F Rotifer Copepod Filament Treatment 100 cells appeared to share a gelatinous matrix. The matrix itself did not stain and was not observed under the light microscope, however the cells adhered to the settling chambers with a particularly vexing tenacity, and a gelatinous sheath covering the single cells was inferred. Because a bloom of such magnitude is likely to influence the filtering behaviour of the metazoan grazers, the densities of Elakatothrix sp. were determined for all treatments on the final day of the Fall experiment. The estimated abundances are shown in Figure 31. Densities were lowest in enclosures with Daphnia and highest in the FILAMENT treatment. The difference in Elakatothrix densities was significant between treatments (Table 13). Homogenous subsets as determined by Tukey HSD multiple comparison tests are shown graphically in Figure 32. Comparison of the treatment means indicated that the FILAMENT treatment had significantly larger Elakatothrix densities than the six enclosures containing Daphnia. Elakatothrix in the DAPHNIA treatment was also significantly lower than the COPEPOD treatment. 4.5 Filament effects - A re-analysis In the Fall experiment, the Fall DAPHNIA - DAPHNIA+F and Fall ROTIFER - FILAMENT treatment pairs were designed, independent of the seasonal comparison, to detect evidence of mechanical interference on Daphnia grazing and to determine the potential impact on microbial food webs. The comparison of the DAPHNIA and DAPHNIA+F treatments explores this question, while the comparison of the FILAMENT and ROTIFER treatment effects can be examined to detect any enhancement of the microbial food web due to inhibition of microbial grazers and/or the availability of increased surface area for microbial attachment. 101 Accordingly, these four treatments were re-analyzed together (excluding the c o p e p o d treatment as there was no C O P E P O D + F treatment to complete the design). Again, repeated measures A N O V A was employed to examine differences in bacterial abundance, while the ciliate, rotifer, and green algae abundances were compared using A N O V A for a single (final) sampling date. These analyses are similar to those described above, the only difference being the exclusion of samples from Fall c o p e p o d enclosures. The exclusion of the c o p e p o d treatment from the statistical analysis resulted in homogenous variances for the entire (5 date) bacteria data set. Thus the problems encountered in the original analysis (violation of the A N O V A assumptions) are not an issue in the filament/no filament comparisons. The results are given in Tables 14a, b and c; the main results are of course similar to those obtained in the initial analysis above. Post-hoc multiple comparisons were made using the Tukey HSD procedure. No comparisons before the November 2 n d sampling date were significant. The results from the November 2 n d and November 4 t h sampling dates are given in Table 14d. On November 2 n d, the d a p h n i a treatment is distinct from the r o t i f e r and f i l a m e n t treatments, and though not nominally different from the D A P H N I A + F treatment, p=0.054, a contrast which had become fully significant by the November 4 t h sampling date. By the final day of the experiment, all four treatments are significantly different. In the initial analysis (which included the Fall c o p e p o d treatment), the need to compare five treatments and the increased inequalities in variance, rendered the test too low in power to detect the more subtle effect of glass fibre filament addition. An examination of Figure 21 illustrated clearly that the effect of Daphnia in enclosures produced a much more pronounced enhancement of bacterial abundance; the 102 Table 14a. Repeated measures A N O V A of natural log-transformed bacterial abundance, comparing the Daphnia, Daphnia+F, Filament and N o Grazer treatments from the F a l l Experiment, 5 dates (°== .05) Within-subject effects Sum of Squares df M e a n Square F Sig . Date 1.431 4 .358 13.354 <.001 Date * Treatment 2.639 12 .220 8.209 <.001 Error (Date) .857 32 .02679 Table 14b. Repeated measures A N O V A of natural log-transformed bacterial abundance , comparing the Daphnia, Daphnia+F, Filament and N o Grazer treatments from the Fa l l Experiment, 5 dates (°== .05) Between-subject Effects Sum of Squares df M e a n Square F Sig . Treatment .906 3 .302 24.170 <.001 Error .09991 8 .01249 Table 14c. Levene's test of equality of error variances: tests the nul l hypothesis that the error variance of the natural log-transformed bacteria abundance in the enclosures is equal for the Daphnia, Daphnia+F, Filament and N o Grazer treatments. Sample date F df 1 df 2 Sig . October 18 .885 3 8 .489 October 25 3.905 3 8 .055 October 30 .890 3 8 .487 November 2 1.346 3 8 .327 November 4 3.818 3 8 .058 103 Table 14d. Tukey H S D multiple comparison tests between the bacterial abundances in the Daphnia, Daphnia+F, Filament and Rotifer enclosures in the Fa l l experiment. Repeated measures A N O V A results are found in Table 14a-c. Bacterial abundances have been log-transformed. Treatment comparisons which were also significant in an analysis which included the F a l l Rotifer treatment are shown in italic type. Date Treatment (i) Treatment (i) M e a n difference ( i - i ) S tandard E r r o r S i g . November 2 Daphnia Filament .2901 ,078 .024 Daphnia+F -.2456 .078 .054 Rotifer .2926 .078 .023 Filament Daphnia+F -.5357 .078 .001 Rotifer .002460 .078 1 Daphnia+F Rotifer .5382 .078 .001 November 4 Daphnia Filament .5937 .071 <.001 Daphnia+F -.2778 .071 .019 Rotifer .8358 .071 <.001 Filament Daphnia+F -.8715 .071 <.001 Rotifer .2421 .071 .038 Daphnia+F Rotifer 1.1136 .071 <.001 104 increased abundance due to "filaments" can be seen to be smaller. But whether Daphnia were present or not, the addition of glass fibre filaments always resulted in a higher bacterial abundance than would otherwise be observed. The pattern seen in other response variables is less clear. Though there is a trend for the treatment mean density of ciliates, rotifers and Elakatothrix sp. to be higher in treatments which received filaments than in the corresponding treatment without filaments, in no case are the DAPHNIA-DAPHNIA+F and FILAMENT-ROTIFER pairs significantly different with regard to these variables (results not shown). At the level of the individual enclosures, most commonly two out of three enclosures followed this trend. 105 5. Discussion 5.1 Zooplankton biomass As expected, there were increases in zooplankton biomass within treatments over the course of the experiment, as the zooplankton populations reproduced in enclosures. The Daphnia treatments attained high population densities, despite the fact that initial densities stocked were low to moderate when compared to similar enclosure studies (Brett et al. 1994; Jurgens et al. 1994a). As there were no Daphnia in Pond 13, the animals had to be harvested from a nearby pond where the food web was dissimilar. This source environment was probably low in food; some of the Daphnia in Library Pond exhibited ephippia, the pond was heavily shaded, and the food web was likely detritus-driven. Once these Daphnia were released into the comparatively lush environment of the enclosures, their growth and reproductive rates (inferred from population growth and changes in size distribution) were high. Based on an estimate of population size at the midpoint of the experiment (data not shown), the Daphnia populations had reached their final abundance levels by the midpoint of the Summer experiment. Though Daphnia densities were high, much higher population sizes, with pronounced effects on bacterial abundance, have been recorded after lake colonization by Daphnia (Jurgens et al. 1994b). Unlike Daphnia, Bosmina was initially present in Pond 13, but at low abundances relative to S. oregonensis and D. brachyurum. The 120 Bosmina L" 1 added to enclosures was higher than ambient density in the Pond, and population increase was modest. Individual Bosmina showed increases in biomass and some reproduction did occur, but population 106 growth was not so large as that observed for Daphnia. The Bosmina populations were apparently close to equilibrium density in the enclosures. The final copepod biomass in Summer enclosures was calculated (using the species-and instar-specific regressions given in Culver et al. 1985) to be smaller than that of Daphnia or Bosmina. The biomass of copepods in the Fall COPEPOD treatment was almost negligible, and I consider the Fall COPEPOD and Fall ROTIFER treatments to be effectively the same. Given the large differences in body size and growth rates of the various grazers, equalizing biomasses is not possible on an experimental time scale that allows reproduction and growth to occur. In evaluating the species-specific effects of the grazers, comparisons of population filtering capacities are much more instructive, and are discussed below. 5.2 Effectiveness of treatments Examination of the zooplankton filtering capacities at the end of both experiments indicated that the DAPHNIA and BOSMINA treatments were successfully imposed as intended, and with comparable grazing pressures between treatments. Estimation of the Summer and Fall DAPHNIA population filtering capacities indicated enclosure turnover times of less than three days. The Summer BOSMINA treatment had an estimated enclosure turnover time only slightly larger, when flagellates were considered the "reference" prey. Bosmina!s estimated filtering capacity on bacteria was much lower, but the Summer DAPHNIA and BOSMINA treatments differ in this respect due to contrasts in feeding behaviour and not due to biomass differences between treatments. This is also true for S. oregonensis in the Summer COPEPOD treatment; while the estimated population clearance rates on flagellates were low for this 107 treatment, when feeding on ciliates the copepods could be expected to clear the entire enclosure in less than two days. A l l of the macrozooplankton treatments had the potential to exert substantial grazing pressure on the microbial food web by the end of the experiments. Their different impacts, both direct and indirect, are due to species-specific differences in feeding behaviour. It is also likely that some of the indirect impacts of each species are a result of species-specific differences in nutrient recycling (excretion), but that possibility was not addressed in this study. The time span of the experiment allowed rotifer populations to grow in all treatments except those containing Daphnia. The treatments imposed resulted in very different grazer communities between enclosures, but only in the Daphnia enclosures were the "single species" treatments maintained. The ubiquitous presence of rotifer populations in all the non-Daphnia treatments does not diminish their comparative value, however. Daphnia can reduce rotifer populations in lakes (Neill 1984); this has not been observed for Bosmina or S. oregonensis. In each of these treatments, the crustacean grazers have the potential for higher grazing rates than the rotifers. Rotifer populations alone are not expected to exert a strong direct influence through grazing microbial food webs, though they may have important indirect impacts on nutrient cycling (Arndt 1993). The presence/absence of rotifers in experimental enclosures is rightly considered an indirect effect of the larger metazoan grazers present, and it is the sum of both direct and indirect impacts that is of interest in this study. The Summer "rotifer" treatment, though not successful in enhancing rotifer abundance above naturally occurring levels, functioned effectively as the C O P E P O D treatment by the end of the experiment. Given the failure of stocked S. oregonensis populations to thrive after 108 experimental manipulation, it is fortuitous, in terms of the experimental design, that S. oregonensis nauplii fared much better than the adults. The NO GRAZER-turned-ROTlFER treatment had been intended to function as a microbial grazer community, but by the time treatment effects began to appear in the other enclosures, the grazer community in the "NO GRAZER" enclosures was rotifer-dominated. In an experiment performed by Brett et al. (1994), each of their treatments also contained rotifer populations, but at a somewhat lower density than I observed in my studies. Their "removal" treatment is equivalent to my Summer ROTIFER treatment. Brett et al. (1994) considered their removal treatment to be "grazer-free" despite rotifer abundances of approximately 75 + 44.7 individuals L" 1. Comparably, mean rotifer abundance in my Summer ROTIFER treatment was 176 ± 80 individuals L" 1. Given the nature of sampling error for rotifer counts, these abundances are not appreciably different (Ruttner-Kolisko 1977). The microbe-only community structure intended for my NO GRAZER treatment exists naturally only in Antarctic lakes. Though such a treatment would have provided an interesting contrast to the grazer treatments, its loss from the design does not reduce the generality of the results. 5.3 Daphnia-mtifer interactions The interactions of macrozooplankton and rotifers can have an indirect influence on the microbial food web. Daphnia virtually excluded rotifers from the enclosures in both the Summer and Fall experiments. Daphnia suppression of rotifer populations is well known from both laboratory (Burns and Gilbert 1986a, b; Maclssac and Gilbert 1991) and field experiments (Neill 1984; Wickham and Gilbert 1991). Rotifer abundances can be 109 suppressed by exploitative or interference competition, or some combination of both (Gilbert 1988a). Daphnia pulex has been shown to interfere with Keratella cochlearis by catching a rotifer in its feeding current and drawing it into the carapace. The rotifer may be rejected immediately with little effect, or drawn up the food grove towards the mouth, with retention time increasing the probability of lethal effects to the rotifer and occasionally resulting in its ingestion (Burns and Gilbert 1986a). Daphnia can also suppress Keratella by exploitative competition (Maclssac and Gilbert 1991). Unlike K. cochlearis, at least one species of Polyarthra is able to escape capture by Daphnia (Gilbert 1988b). It has been suggested that this response of Polyarthra is affected by container size, with Polyarthra being suppressed by Daphnia in small enclosures but not in larger ones (Sarnelle 1997). One possible explanation for this is that long incubation times may increase the probability of encounter for Daphnia and Polyarthra in small enclosures (Wickham and Gilbert 1991), resulting in a stronger measured effect. However, my enclosures were much larger than the glass jars used previously (Wickham and Gilbert 1991), and it is possible that the high grazing pressure of Daphnia resulted in both exploitative and interference competition with Polyarthra, as Daphnia heavily grazed all edible algae in the enclosures. Polyarthra was greatly suppressed by Daphnia in both the Summer and Fall experiments. 110 5.4 Bosmina-rotifer interactions The B O S M I N A treatment, though successful in maintaining a fiosmma-dominated community, resulted in an interesting but difficult-to-interpret enhancement of rotifer abundances. It would be expected that Bosmina longirostris'? body size (maximum ~ 450 pm in length) is too small to allow interaction with rotifers by direct interference in the same manner as Daphnia (Wickham and Gilbert 1991). Given that K. cochlearis, P. vulgaris and Bosmina may all compete for the same preferred food (small flagellates), it is difficult to explain why rotifer abundances would be highest in the presence of Bosmina. It suggests that nutrient cycling or other indirect effects of the presence of Bosmina longirostris may enhance both the microbial food web and the microzooplankton. 5.5 Productivity and nutrient cycling Though bacterial production was not measured in this study, the model outlined in Table 1 predicts a high ratio of bacterial production to primary production under Daphnia grazing (Jurgens 1994; see also Jeppesen et al. 1992) . Daphnia may decrease bacterial abundance by cropping bacteria cells directly, but grazing releases algal carbon and recycles potentially limiting nutrients, both of which can stimulate bacterial growth (Olsen et al. 1986, Jurgens 1994). Low levels of grazing may allow the indirect benefits to be of greater magnitude than the negative impact of direct grazing. However, algae were grazed to very low levels in my Summer D A P H N I A enclosures. I suspect that in the Summer D A P H N I A enclosures bacterial productivity was negatively affected by the reduction in algal biomass (and the concomitant reduction of available carbon substrates). The contention that moderate 111 grazing may enhance productivity (Sterner 1986) has yet to be definitively tested for bacterioplankton. Ultimately, however, bacterial production in a Daphnia-dominated system may be channelled to higher trophic levels, while in food webs dominated by small zooplankton, most of the bacterial production is respired within the microbial food web. Enhancement of bacterial production and turnover by the indirect effects of grazing is of little importance to the classical lake food web if bacterial carbon does not pass into the zooplankton via direct pathways. Even if Daphnia were to decrease bacterial productivity in absolute terms, it is able to convert bacterial production into metazoan biomass, while smaller cladocerans and copepods cannot. 5.6 Daphnia and bacteria in Summer The effect of the Summer DAPHNIA treatment relative to the BOSMINA and Summer COPEPOD treatments upholds the model of Daphnia interactions in microbial food webs (Table 1). Daphnia was able to graze down all the algae in the enclosure and hold bacterial abundance low through both direct and indirect effects. This stands in contrast to the Summer COPEPOD and BOSMINA treatments, where bacterial abundances were relatively high. The Summer DAPHNIA treatment had a bacteria standing stock similar to the Summer ROTIFER treatment, but in no other way were the food webs similar. There were grazable algae remaining in Summer ROTIFER enclosures, but little other than Volvox colonies remained in the Summer DAPHNIA enclosures. Ciliate densities in the Summer ROTIFER enclosures were significantly higher than those in the DAPHNIA treatment, and rotifers were all but excluded from the Summer DAPHNIA enclosures. The ROTIFER treatment has the potential 112 for high protistan grazing pressure on bacteria. Heterotrophic flagellates (main bacterivores) were likely grazed by the rotifer community in the ROTIFER enclosure, but neither Keratella nor Polyarthra feeds on ciliates (Buikema et al. 1978, Gilbert and Bogdan 1981, Arndt 1993). However, the absence of small zooplankton from ROTIFER enclosures such as Bosmina and copepods may also have had an indirect impact on the bacteria by decreasing nutrient recycling. The absence of macrozooplankton grazers probably denies bacteria the algal carbon made available to them by sloppy feeding. My results indicate that the loss of the positive indirect effects of small zooplankton on the microbial food web has the same consequence for bacterial abundance as high Daphnia grazing. In the Summer DAPHNIA enclosures, algae were grazed down to such low levels that the Daphnia were able to consume a substantial portion of the bacterial standing stock, and also deny bacteria the substrates (algal exudates) needed for growth. 5.7 Daphnia and bacteria in Fall In the Fall experiment, the difference between the high and low bacteria densities resulting from treatment effects was even more pronounced than that seen in the summer. However, in the case of DAPHNIA treatments, the effect was opposite to that seen in the Summer experiment. As the Daphnia biomass (and filtering capacity) were the same in the Summer and Fall DAPHNIA treatments, this difference in outcome does not result from a difference in the grazer community. The high bacterial abundance in Fall suggests that the pathway for Daphnia's direct effects on bacteria had been inhibited. Daphnid?, impact on ciliates and rotifers remained similar to the Summer treatment. The observed bloom of 113 Elakatothrix, present in Fall enclosures but not Summer, likely altered Daphnid? filtering of the bacterial size fraction. Elakatothrix densities were lowest in the Daphnia enclosures, which is indicative of some cropping of the algae. Saturation of Daphnid? feeding by high food concentration results in a plateau of ingestion rate and a decline in overall filtering rate (Lampert 1987a). Daphnia feeds inefficiently on bacteria and is only able to ingest the largest fraction of the available cells (Brendelberger 1991). The bloom of algae likely saturated Daphnid? ingestion rates. This may have occurred if the algae were a good food source, and if not, the interference of so many low quality food particles would also reduce filtering rates. Daphnid? impact on the microbial food web can thus be heavily influenced by the bottom up mechanisms which drive nutrient regimes and the development of algal blooms. My results indicate that Daphnia will always exert some kind of top-down control of microbial food webs, but the outcome may be either indirect enhancement of bacterial density or direct suppression of bacterial abundance. The existence of a large algal bloom in the Fall enclosures may also indicate an increase in nutrient availability. Both algae and bacteria require dissolved nutrients for growth, and actively compete for them (Currie and Kalff 1984). Bacteria are also dependant on dissolved carbon substrates, which may be increased in the presence of an algal bloom (through algal exudation/lysis and "sloppy feeding" of grazers). 114 5.8 Other zooplankton and bacteria Unfortunately, Bosmina's impact on bacteria cannot be assessed in the presence of the Elakatothrix bloom, as the animals were not available for experimental collection in Fall. The evidence regarding Bosmina is specific to summer only, and as predicted, bacterial abundance was high in the Bosmma-dominated community. This is consistent with evidence that Bosmina does not graze bacteria directly (Bogdan and Gilbert 1982, Hart 1996). Though Bosmina could potentially graze bacterial predators in the < 20 urn size fraction (heterotrophic nanoflagellates), this cannot be determined from the data available. As stated above, Bosmina's effect on ciliate abundance was moderate, despite its potentially high clearance rates for ciliates (Sanders and Wickham 1993). Rotifer abundances were higher in the Bosm ma-dominated community than in the R O T I F E R communities, and yet Bosmina has been shown to selectively graze the small flagellates which are also the preferred food of K. cochlearis and P. vulgaris (Bogdan and Gilbert 1982). Enhancement of bacterial abundance is expected in a Bosmma-dominated community (Table 1), and it would appear that protistan and rotifer populations benefit from Bosminds presence as well. Bosmina's dual feeding mode allows it to co-exist with Daphnia rather than becoming excluded by exploitative competition (DeMott 1982). Its relationships with microzooplankton competitors may also be complex, but the design of this study does not allow this possibility to be fully assessed. The high bacterial abundances seen in the Summer C O P E P O D treatment are also in keeping with the predictions of the Daphnia vs. small zooplankton model (Table 1), though the pathways for the food web interactions are different. When compared to the R O T I F E R treatment, enclosures with copepods demonstrated the predicted high bacteria densities, while 115 the rotifer dominated enclosures did not. Rotifer-dominated zooplankton communities are maintained by planktivorous fish predation on macrozooplankton in some lakes, but more often an abiotic factor such as pH excludes other competing zooplankton (Arndt 1993). The effects of rotifers on microbial food webs have been reviewed, but their impact is seldom studied in isolation (Arndt 1993). Though K. cochlearis is capable of grazing bacteria, P. vulgaris is not (Sanders et al. 1989), and in both Summer and Fall enclosures, the rotifer dominated communities expressed low bacteria densities. This leads me to suspect that it is the combination of both 1) unfettered protistan grazing pressure and 2) the loss of the positive indirect effects of macrozooplankton grazing (nutrient recycling and sloppy feeding) which dictated bacterial abundance in the R O T I F E R treatments. In the absence of direct grazing measurements and nutrient measurements, this conclusion is purely speculative. My results, in general, demonstrate the high bacterial abundances predicted for small zooplankton-dominated communities, but microzooplankton and microbial grazers do not fit this pattern in isolation. Usually the term "small zooplankton dominated" is used to encompass mixed copepod, small cladoceran and rotifer communities, but the data in this study indicate that rotifers are not equivalent to the macrozooplankton groups in their impact on the microbial food web. 5.9 Zooplankton-ciliate interactions The interaction of Daphnia and ciliates has been less well studied than that of Daphnia and rotifers, but the evidence points to the same general conclusions. Daphnia are able to suppress ciliates by both interference and exploitative competition (Neill 1984, Gilbert 1988a, 116 Wickham and Gilbert 1991). Unlike rotifers, ciliates can also form a nutritive component of Daphnid? diet, though as for algae, the nutritive value of a particular ciliate species may differ for the various zooplankton taxa (DeBiase et al. 1990, Sanders et al. 1989, Sanders and Wickham 1993 ). Many ciliates have escape responses which would also dictate species-specific vulnerability to zooplankton predators (Wickham and Gilbert 1991, Jack and Gilbert 1993, Sarnelle 1997). Daphnia pulex has been shown to depress ciliate abundance while in the same study Bosmina longirostris did not (Wickham and Gilbert 1991). Though ciliate densities in the BOSMINA treatment in this study were somewhat less than that found in the rotifer dominated enclosures, the ciliate abundance under Bosmina was higher than in the DAPHNIA treatments. This occurred despite the known potential of Bosmina to feed on ciliates (Sanders and Wickham 1993). The Summer and Fall DAPHNIA treatments had the lowest ciliate abundances observed in the study, while the DAPHNIA+F treatment (grazing interference) did not exhibit this suppression to the same degree. The Summer COPEPOD treatment also had relatively low ciliate abundances; Skistodiaptomus oregonensis has been shown to thrive on a diet of ciliates in the laboratory (Sanders et al. 1996), and likely preyed heavily on ciliates in the enclosures. The Elakatothrix bloom which probably depressed Daphnid? grazing on bacteria did not substantively alter Daphnid? effect on ciliate abundance. Daphnid? grazing of ciliates is likely to be a function of encounter rate and the defensive mechanisms of the ciliate (Jack and Gilbert 1993). Daphnid? clearance rate on ciliates is more a function of encounter rate than particle retention, and may not be affected by mechanical interference to the same extent as 117 that for bacteria. A similar enhancement of bacterial abundance by the indirect pathway of predation on ciliates was also seen in the S. oregonensis-dominated community of the Summer C O P E P O D treatment. 5.10 Algal blooms and grazing interference One major objective of this study was to assess seasonal differences in the response of bacterial abundance to zooplankton grazing. By repeating treatments in time as well as replicating within an experiment, it is possible to assess the general applicability of the results. While the rotifer-dominated communities had similar effects on bacterial abundance across seasons, the impact of Daphnia varied with seasonal differences in algal abundance and diversity. Algal blooms are a common feature of lake phytoplankton dynamics. Late summer algal communities often exhibit increased abundance of inedible algae in response to zooplankton grazing, while successional and grazer induced shifts in dissolved nutrient ratios may favour blooms of filamentous cyanobacteria or indigestible gelatinous green algae (Sommer et al. 1986). Though it is not possible to conclusively determine the cause of the Fall algal bloom observed in the experimental enclosures, rainfall may have provided substantial nutrient inputs to the ponds and the enclosures. The precipitation-weighted average nitrogen content of rainfall near the University of British Columbia , measured in 1991 in the Georgia Basin (Strait of Georgia), has been estimated at 17 ± 2.5 u M [N0 3+NH 4] (Mackas and Harrison 1997). Nutrient inputs to the enclosures via rainfall may have precipitated the observed bloom of Elakatothrix sp., as rainfall occurred almost daily during 118 the fall experiment. Irrespective of its origin, the occurrence of a Fall Elakatothrix bloom in the experimental enclosures allows the effects of zooplankton grazing on bacterial abundance to be tested for a food web configuration not explicitly addressed by the model in Table 1. Though Daphnids feeding responses to food concentration and food quality have been well characterized in the lab (Lampert 1987a), this wide range of potential responses is often overlooked in both predictive models and field experiments. As my results suggest, algal blooms which alter Daphnids grazing behaviour may allow the bacterioplankton to escape top-down control by macrozooplankton. Algal filaments also enhance bacterial growth during algal senescence and lysis. Many species of cyanobacteria can inhibit Daphnia grazing. Often they are toxic to zooplankton, and filamentous forms can mechanically inhibit grazing (Lampert 1987b). Model filaments have been used to investigate this mechanical effect, but their success has been somewhat limited (Webster and Peters 1978). More commonly, natural filaments have been used to illustrate the mechanics of grazing inhibition (Lampert 1987b). The use of natural filaments to determine mechanical interference of Daphnia grazing on bacteria is problematic; the filaments may release organic substrates as they decay, and enhance bacterial growth still further while the zooplankton grazing is inhibited. By using a model filament with no nutritive value to the Daphnia or the bacteria, I was able to detect the effect of mechanical interference on Daphnia grazing bacteria, without providing additional substrates for bacteria growth. However, the significantly enhanced bacterial abundance in the F I L A M E N T treatment, indicates that the physical presence of suspended filaments can enhance bacterial abundance independent of nutritional effects. 119 The addition of filaments in the absence of large zooplankton ( F I L A M E N T treatment) provided suspended particles which enhanced microbial growth. Bacteria are known to attach to suspended organic particles in both marine and freshwater environments (Simon 1987). The productivity and cell size of attached bacteria are much greater than free living forms (Kirchman 1983, Simon 1987). Much of this productivity increase is thought to be provided by increased substrate availability in the vicinity of flocculent organic matter. However, this study demonstrates that bacterial growth or biomass can be stimulated by increased (inorganic) surface area available for attachment. Adding glass fibre filaments increased the spatial complexity and effective "surface area" of the enclosure environment. It is possible that algal blooms function in a similar way. Not only do filamentous algae inhibit zooplankton grazing, they also provide a physical matrix for enhanced microbial activity. This appears to be the case in the F I L A M E N T treatment, where bacterial abundance was significantly higher than in the Fall R O T I F E R treatment (filament free). There was a slight trend for other components of the F I L A M E N T food web (ciliates, rotifers and Elakatothrix) to be higher as well, but not significantly so. The results of the Fall experiment suggest, that in environments where suspended inorganic particles interfere with Daphnia grazing (Kirk and Gilbert 1990, Kirk 1991), bacteria densities may be enhanced. The D A P H N I A + F treatment demonstrates this enhancement, while the increased bacterial abundance in the F I L A M E N T treatments suggests that the increase is due to both grazing inhibition and increased particle surface area for microbial attachment. The increased spatial complexity generated by suspended particles has a measurable impact on microbial processes. 120 5.11 Comparison to other studies The Summer experiment agrees well with results from other enclosure studies of bacterial abundance. Bacterial abundance was low in the D A P H N I A treatment as predicted by the model outlined in Jurgens (1994). Bacterial abundance was high under the B O S M I N A treatment, in accordance with the results obtained by Jeppesen et al. (1992) for a community dominated by B. longirostris, and Geertz-Hansen et al. (1987) for B. coregoni-dommat&d enclosures. The high bacterial abundance I observed in the Summer C O P E P O D treatment is supported by inspection of the results of Brett et al. (1994), who found a significant increase in bacterial abundance in small enclosures containing Diaptomus novamexicanus relative to those dominated by Daphnia rosea. In previous studies comparable to mine, the impact of rotifers has not been examined in the absence of other small zooplankton. The closest example is the "removal" treatment of Brett et al. (1994) which had somewhat lower rotifer abundances than my "rotifer-dominated" enclosures. None of the zooplankton treatments in Brett et al. (1994) had bacterial densities significantly different from the "removal" treatment, where bacterial abundance was intermediate between the Daphnia and small zooplankton treatments. In contrast, the enhancement of bacterial abundance in the Fall D A P H N I A treatments goes against the trend in other studies towards reduced abundance. As most other studies took place in the Summer, it is possible that seasonality may play a greater role in bacteria-zooplankton interactions than has been investigated to date. Experiments conducted in one season cannot necessarily be extrapolated to other time periods (Brett et al. 1994). A wider 121 range of food web states needs to be examined before an all-encompassing model of Daphnia's impact can be validated. My study confirms the results of previous studies for the Summer food web condition, but the opposite effect observed in the Fall indicates that the impact of Daphnia on bacterioplankton has not yet been fully characterized. Daphnia can have a large impact on bacteria where it is abundant, but the balance of its direct and indirect impacts on the microbial food web may differ seasonally. The presence of Daphnia can be a strong predictor of bacterial abundance. However, knowledge of other factors affecting Daphnia grazing, such as food quality, quantity and/or the presence of inhibitory algal blooms must also be factored into any predictive model. While the magnitude of Daphnia's impact on microbial food webs is.often large, the effect on bacterial abundance is not always negative. 122 References Arndt, H . (1993) Rotifers as predators on components of the microbial web (bacteria, heterotrophic flagellates, ciliates) - a review. Hydrobiol. 255/256: 231-246. Azam, F, T. Fenchel, J.G. Field, J.S. Gray, L A . Meyer-Reil, F. Thingstad. (1983) The ecological role of water-column microbes in the sea. Mar. Ecol. Prog. Ser. 10:257-263. Bengtsson, J. (1987) Competitive dominance among Cladocera: Are single-factor explanations enough? Hydrobiol. 145: 245-257. Bennet, S.J., R.W. Sanders and K.G. Porter. (1990) Heterotrophic, autotrophic, and mixotrophic nanoflagellates: Seasonal abundances and bacterivory in a eutrophic lake. Limnol. Oceanogr. 35: 1821-1832. Berninger, Ulrike-G., B.J. Finlay and P. Kuuppo-Leinikki. (1991) Protozoan control of bacterial abundances in freshwater. Limnol. Oceanogr. 36: 139-147. Bird, D.F. and J. Kalff (1984): Empirical relationships between bacterial abundance and chlorophyll concentration in fresh and marine waters. Can. J. Fish. Aquat. Sci. 41: 1015-1023. Bleiwas, A . H . and P.M. Stokes. (1985) Collection of large and small food particles by Bosmina. Limnol. Oceanogr. 30: 1090-1092. 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. Limnol. Oceanogr. 27: 918-934. Bogdan, K .G. and J.J. Gilbert. (1987) Quantitative comparison of food niches in some freshwater zooplankton. A multi-tracer-cell approach. Oecologia 72: 331-340. Bogdan, K.G. , J.J. Gilbert and P.L. Starkweather. (1980) In situ clearance rates of planktonic rotifers. Hydrobiol. 73: 73-77. 123 Boraas, M.E., K.W. Estep, P.W. Johnson and J.McN. Sieburth. (1988) Phagotrophic phototrophs: The ecological significance of mixotrophy. J. Protozool. 35: 249-252. Bottrell, H. , A . Duncan, Z .M. Gliwicz, E. Grygierek, A . Herzig, A . Hillbricht-Ilkowska, H . Kurasawa, P. Larsson and T. Weglenska. (1976) A review of some problems in zooplankton production studies. Norw. J. Zool. 24: 419-456. Brendelberger, H . (1985) Filter mesh size and retention efficiencies for small particles: Comparative studies with Cladocera. Archiv. Hydrobiol. Beih. Ergebn. Limnol. 21: 135-146. Brendelberger, H . (1991) Filter mesh size of cladocerans predicts retention efficiency for bacteria. Limnol. Oceanogr. 36: 884-894. Brett, M.T., K. Wiackowski, F.S. Lubnow, A . Mueller-Solger, J.J. Elser and C.R. Goldman. (1994) Species-dependent effects of zooplankton on planktonic ecosystem processes in Castle Lake, California. Ecology 75: 2243-2254. Brinch-Iversen, J. and G.M. King. (1990) Effects of substrate concentration, growth state, and oxygen availability on relationships among bacterial carbon, nitrogen and phospholipid phosphorus content. FEMS Microbiol. Ecol. 74: 345-356. Buikema, A.L . , J.D. Miller and W.H. Yongue. (1978) Effects of algae and protozoans on the dynamics of Polyarthra vulgaris. Verh. Internat. Verein. Limnol. 20: 2395-2399. Burns, C.W., and J.J. Gilbert. (1986a) Direct observations of the mechanisms of interference between Daphnia and Keratella cochlearis. Limnol. Oceanogr. 31: 859-866. Burns, C.W., and J.J. Gilbert. (1986b) Effects of daphnid size and density on interference between Daphnia and Keratella cochlearis. Limnol. Oceanogr. 31: 848-858. Burns, C.W. and J.J. Gilbert (1993). Predation on ciliates by freshwater calanoid copepods: rates of predation and relative vulnerabilities of prey. Freshwater Biology 30: 377-393. 124 Butler, N .M. , Suttle,CA. and W.E. Neill. (1989) Discrimination by freshwater zooplankton between single algal cells differing in nutritional status. Oecologia 78: 368-372. Carpenter, S.R. J.F. Kitchell and J.R. Hodgson. (1985) Cascading trophic interactions and lake productivity. Bioscience 35: 634-639. Carpenter, S.R., J.F. Kitchell, J.R. Hodgson, P.A. Cochran, J.J. Elser, M . M . Elser, D . M . Lodge, D. Kretchmer, X . He and C.N. von Ende. (1987) Regulation of lake primary productivity by food web structure. Ecology 68: 1863-1876. Christoffersen, K., B. Reimann, L.R. Hansen, A . Klysner, H . Soernsen. (1990) Qualitative importance of the microbial loop and plankton community structure in a eutrophic lake during a bloom of cyanobacteria. Microb. Ecol. 20: 253-272 Christofferson, 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. Limnol. Oceanogr. 38: 561-573. Currie, D.J. (1990) Large-scale variability and interactions among phytoplankton, bacterioplankton and phosphorus. Limnol. Oceanogr. 35: 1437-1455. Currie, D.J. and J. Kalff. (1984) A comparison of the abilities of freshwater algae and bacteria to aquire and retain phosphorus. Limnol. Oceanogr. 29: 298-310. Currie, D.J., E. Bentzen and J. Kalff. (1986) Does algal-bacterial phosphorus partioning vary among lakes? A comparative study of orthophosphate uptake and alkaline phosphatase activity in freshwater. Can. J. Fish. Aquat. Sci. 43: 311-318. DeBiase, A.E. , R.W. Sanders and K.G. Porter. (1990) Relative nutritional value of ciliate protozoa and algae as food for Daphnia. Microb. Ecol. 19: 199-210. del Giorgio, P.A. and G. Scarborough. (1995) Increase in the proportion of metabolically active bacteria along gradients of enrichment in freshwater and marine plankton: implications for estimates of bacterial growth and production rates. J. Plankton Res. 17: 1905-1924. 125 DeMott, W.R. (1982) Feeding selectivities and relative ingestion rates of Daphnia and Bosmina. Limnol. Oceanogr. 27: 518-527. Demott, W.R. and W.C. Kerfoot. (1982) Competition among cladocerans: nature of the interaction between Daphnia and Bosmina. Ecology 63: 1949-1966. Demott, W.R. and M.D. Watson. (1991) Remote detection of algae by copepods: responses to algal size, odors and motility. J. Plankton Res. 13: 1203-1222. Dodson, S.I. (1974) Zooplankton competition and predation: an experimental test of the size-efficiency hypothesis. Ecology 55: 605-613. Ducklow, H.W., D.A. Purdie, PJ.LeB.Williams and J.M. Davies. (1986) Bacterioplankton: A sink for carbon in a coastal marine plankton community. Science 232: 865-867. Dumont, H.J., I. Van de Velde and S. Dumont. (1975) The dry weight estimate of biomass in a selection of Cladocera, Copepoda and Rotifera from the plankton, periphyton and benthos of continental waters. Oecologia 19: 75-97. Fenchel, T. (1982) Ecology of heterotrophic microflagellates. IV. Quantitative occurrence and importance as bacterial consumers. Mar. Ecol. Prog. Ser. 9: 35-42. Fuhrman, J.A. and F. Azam. (1982) Thymidine incorporation as a measure of heterotrophic bacterioplankton production in marine surface waters: Evaluation and field results. Mar. Biol. 66: 109-120. Fuhrman, J.A., K. McCallum and A . A . Davis. (1992) Novel major archaebacterial group from marine plankton. Nature 356: 148-149. Geertz-Hansen, O., M . Olesen, P. Koefoed Bjornsen, J. Brenner Larsen and B. Riemann. (1987) Zooplankton consumption of bacteria in a eutrophic lake and in experimental enclosures. Arch. Hydrobiol. 110: 553-563. 126 Gilbert, J.J. (1988a) Suppression of rotifer populations by Daphnia: A review of the evidence, the mechanisms, and the effects on zooplankton community structure. Limnol. Oceanogr. 33: 1286-1303. Gilbert, J.J. (1988b) Susceptibilities of ten rotifer species to interference from Daphnia pulex. Ecology 69: 1826-1838. Gilbert, J.J. and K . G . Bogdan. (1981) Selectivity of Polyarthra and Keratella for flagellate and aflagellate cells. Verh. Internat. Verein. Limnol. 21: 1515-1521. Gliwicz, Z . M . (1990) Food thresholds and body size in cladocerans. Nature 343: 638-640. Gliwicz, Z . M . (1980) Filtering rates, food size selection, and feeding rates in cladocerans -another aspect of interspecific competition in filter-feeding zooplankton. Am. Soc. Limnol. Ocean. Spec. Symp 3: 282-292. Gliwicz, Z . M . and W. Lampert (1993) Body-size related survival of cladocerans in a trophic gradient: an enclosure study. Arch. Hydrobiol. 129: 1-23. Gude, H . (1988) Direct and indirect influences of crustacean zooplankton on bacterioplankton of Lake Constance. Hydrobiol. 159: 63-73. Gude, H . (1990) The role of grazing on bacteria in plankton succession. Chap. 9. In: Phytoplankton Ecology: Succession in plankton communities. 1st ed. (Ed: Sommer,U.) Springer-Verlag, Berlin Heidelberg, 337-364. Gude, H . (1991) Bacterial production and the flow of organic matter in Lake Constance. Chap. 25. In: Large Lakes, Eds. Tilzer and Serruya, Springer-Verlag, New York, 489-502. Hall, D.J., S.T. Threlkeld, C. Burns and P.H. Crowley. (1976) The size-efficiency hypothesis and the size structure of zooplankton communities. Ann. Rev. Ecol. Syst. 7: 177-208. 127 Hardy, F.J., K.S. Shortreed and J.G. Stockner (1986) Bacterioplankton, phytoplankton, and zooplankton communities in a British Columbia coastal lake before and after nutrient reduction. Can. J. Fish. Sci. 43: 1504-1514. Hart, R.C. (1996) Naupliar and copepodite growth and survival of two freshwater calanoids at various food levels: Demographic contrasts, similarities, and food needs. Limnol. Oceanogr. 41: 648-658. Jack, J.D. and J.J. Gilbert. (1993) Susceptibilities of different-sized ciliates to direct suppression by small and large cladocerans. Freshwater Biology 29: 19-29. Jack, J.D. and J.J. Gilbert. (1994) Effects of Daphnia on microzooplankton communities. J. Plankton Res. 16: 1499-1512. Jeppesen, E., O. Sortkjaer, M . Sondergaard and M . Erlandsen. (1992) Impact of a trophic cascade on heterotrophic bacterioplankton production in two shallow fish-manipulated lakes. Arch. Hydrobiol. Beih. Ergebn. Limnol. 37: 219-231. Jurgens, K. and G. Stolpe. (1995) Seasonal dynamics of crustacean zooplankton, heterotrophic nanoflagellates and bacteria in a shallow, eutrophic lake. Freshwater Biology 33: 27-38. Jurgens, K., H . Arndt and K.O. Rothhaupt. (1994a) Zooplankton-mediated changes of bacterial community structure. Microb. Ecol. 27: 27-42. Jurgens, K., J.M. Gasol, R. Massana and C. Pedros-Alio. (1994b) Control of heterotrophic bacteria and protozoans by Daphnia pulex in the epilimnion of Lake Ciso. Arch. Hydrobiol. 131: 55-78. Jurgens, K. (1994) Impact of Daphnia on planktonic microbial food webs- a review. Mar. Mic. Food Webs 8: 295-324. Kerfoot, W.C. and K.L . Kirk. (1991) Degree of taste discrimination among suspension-feeding cladocerans and copepods: Implications for detritivory and herbivory. Limnol. Oceanogr. 36: 1107-1123. 128 Kirchman, D.L. (1993) Statistical Analysis of Direct Counts of Microbial Abundance. Chap. 14. In: Handbook of Methods in Aquatic Microbial Ecology, ed. P.F. Kemp et al., Lewis Publishers, Boca Raton, p.117-120. Kirchman, D.L. (1983) The production of bacteria attached to particles suspended in a freshwater pond. Limnol. Oceanogr. 28: 858-872. Kirk, K . L . (1991) Inorganic particles alter competition in grazing plankton: the role of selective feeding. Ecology 72: 915-923. Kirk, K . L . and J.J. Gilbert. (1990) Suspended clay and the population dynamics of planktonic rotifers and cladocerans. Ecology 71: 1741-1755. Kirk, R.E. (1982) Experimental design: procedures for the behavioral sciences. 2nd ed. Brooks/Cole Publishing Company, Belmont, California. 911 p. Knochel, R. and L.B. Holtby. (1986) Construction and validation of a body-length-based model for the prediction of cladoceran community filtering rates. Limnol. Oceanogr. 31: 1-16. Lampert, W. (1987a) Feeding and nutrition in Daphnia. Mem. 1st. Ital. Idrobiol. 45: 143-192. Lampert, W. (1987b) Laboratory studies on zooplankton-cyanobacteria interactions. New Zealand Journal of Marine and Freshwater Research 21: 483-490. Lampert, W. (1994) Phenotypic plasticity of the filter screens in Daphnia: Adapataion to a low-food environment. Limnol. Oceanogr. 39: 997-1006. Lampert, W., W. Fleckner, H . Rai and B.E. Taylor. (1986) Phytoplankton control by grazing zooplankton: A study on the spring clear-water phase. Limnol. Oceanogr. 31: 478-490. Malley, D.F., S.G. Lawrence, M A . Maclver and W.J. Findlay. (1989) Range of variation in estimates of dry weight for planktonic Crustacea and rotifera from temperate North American lakes. Can. Tech. Rep. Fish. Aquat. Sci. 1666, 49p. 129 MacAuley, E. (1984) The estimation of the abundance and biomass of zooplankton in samples. In: A manual on methods for the assessment of secondary productivity in fresh waters. (Ed. Downing, J.A., Rigler, F.H.). Blackwell Scientific, Boston, 228-265. Maclssac, H.J. and J.J. Gilbert. (1991) Discrimination between exploitative and interference competition between cladocera and Keratella cochlearis. Ecology 72: 924-937. Mackas, D. L. and P. J. Harrison. (1997) Nitrogenous nutrient sources and sinks in the Juan de Fuca Strait/Strait of Georgia/Puget Sound estuarine system: Assessing the potential for eutrophication. Estuarine and Coastal Shelf Science 44: 1-21. Markosova, R. and J. Jezek. (1993) Bacterioplankton interactions with Daphnia and algae in experiemtnal enclosures. Hydrobiol. 264: 85-99. McQueen, D.J., M.R.S. Johannes, J.L. Post,JR and D.R.S. Lean. (1989) Bottom-up and top-down impacts on freshwater pelagic community structure. Ecol. Monog. 59: 289-309 Muller, H. , A . Schone, R .M. Pinto-Coelho, A . Schweizer and T. Weisse. (1991) Seasonal succession of ciliates in Lake Constance. Microb. Ecol. 21: 119-138. Neill, W.E. (1984) Regulation of rotifer densities by crustacean zooplankton in an oligotrophic montane lake in British Columbia. Oecologia 61: 175-181. Olsen, Y. , A . Jensen, H. Reinertsen, K . Y . Borsheim, M . Heldal and A . Langeland. (1986) Dependence of the rate of release of phosphorus by zooplankton on the P:C ratio in the food supply, as calculated by a recycling model. Limnol. Oceanogr. 31: 34-44. Pace, M.L. and J.J. Cole. (1996) Regulation of bacteria by resources and predation tested in whole-lake experiments. Limnol. Oceanogr. 41: 1448-1460. Pace, M.L. , G.B. McManus and S.E.G. Findlay. (1990) Planktonic community structure determines the fate of bacterial production in a temperate lake. Limnol. Oceanogr. 35: 795-808. 130 Pace, M . L . and E. Funke. (1991) Regulation of planktonic microbial communities by nutrients and herbivores. Ecology 72: 904-914. Pennak, R.W. (1989) Fresh-water invertebrates of the United States. 3rd ed. John Wiley and Sons, Toronto. 628 p. Porter, K . G . and Y.S. Feig. (1980): The use of DAPI for identifying and counting aquatic microflora. Limnol. Oceanogr. 25: 943-948. Porter, K.G. , Y.S. Feig and E.F. Vetter. (1983) Morphology, flow regimes, and filtering rates of Daphnia, Ceriodaphnia, and Bosmina fed natural bacteria. Oecologia 58: 156-163. Porter, K.G. , H . Paerl, R. Hodson, M . Pace, J. Priscu, B. Reimann, D. Scavia and J. Stockner. (1988) Microbial interactions in lake food webs. In: Complex Interactions in Lake Communities. (Ed: Carpenter,S.R.) Springer-Verlag, New York, N Y , 209-227. Porter, K . G . (1988) Phagotrophic phytoflagellates in microbial food webs. Hydrobiol. 159: 89-97. Pourriot, R. (1977) Food and feeding habits of Rotifera. Arch. Hydrobiol. Beih. Ergebn. Limnol. 8: 243-260. Psenner, R. and R. Sommaruga. (1992) Are rapid changes in bacterial biomass caused by shifts from top-down to bottom-up control? Limnol. Oceanogr. 37: 1092-1100. Riemann, B. and M . Sondergaard. (1986) Regulation of bacterial secondary production in two eutrophic lakes and in experimental enclosures. J. Plankton Res. 8: 519-536. Romanovsky, Y.E . (1985) Food limitation and life-history strategies in cladoceran crustaceans. Arch. Hydrobiol. 21: 363-372. Ruttner-Kolisko, A . (1977) Comparison of various sampling techniques, and results of repeated sampling of planktonic rotifers. Arch. Hydrobiol. Beih. Ergebn. Limnol. 8: 13-18. 131 Sanders, R.W. and S.A. Wickham. (1993) Planktonic protozoa and metazoa: predation, food quality and population control. Mar. Mic. Food Webs 7: 197-223. Sanders, R.W., C.E. Williamson, P.L. Stutzman, R.E. Moeller, C.E. Goulden and R. Aoki-Goldsmith. (1966) Reproductive success of herbivorous zooplankton fed algal and non-algal food resources. Limnol. Oceanogr. 41: 1295-1305. Sanders, R.W., D.A. Leeper, C.H. King and K . G . Porter. (1994) Grazing by rotifers and crustacean zooplankton on nanoplanktonic protists. Hydrobiol. 288: 167-181. Sanders, R.W. and K. Porter. (1990) Bacterivorous flagellates as food sources for the freshwater crustacean zooplankter Daphnia ambigua. Limnol. Oceanogr. 35: 188-191. Sanders, R.W., K . G . Porter, S.J. Bennett and A.E . DeBiase (1989) Seasonal patterns of bacterivory by flagellates, ciliates, rotifers, and cladocerans in a freshwater planktonic community. Limnol. Oceanogr. 34: 673-687. Sarnelle, O. (1997) Daphnia effects on microzooplankton: comparisons of enclosure and whole-lake responses. Ecology 78: 913-928. Schluter, D. (1994) Experimental evidence that competition promotes divergence in adaptive radiation. Science 266: 798-801. Schoenberg, S.A. (1990) Short-term productivity responses of algae and bacteria to zooplankton grazing in two freshwater lakes. Freshwater Biology 23: 395-410. Schoenberg, S.A. and A.E. Maccubbin. (1985) Relative feeding rates on free and particle-bound bacteria by freshwater macrozooplankton. Limnol. Oceanogr. 30: 10.84-1090. Sherr, E.B., B.F. Sherr and L.J. Albright. (1986) Bacteria: Link or sink? Science 235: 88-89 Sherr, B.F.; E.B. Sherr and J. McDaniel. (1992) Effect of protistan grazing on the frequency of dividing cells in bacterioplankton assemblages. Appl. Environ. Microbiol. 58: 2381-2385. 132 Simon, M . (1987) Biomass and production of small and large free-living and attached bacteria in Lake Constance. Limnol. Oceanogr. 32: 591-607. Sommer, U. , M.Z. Gliwicz, W. Lampert and A. Duncan. (1986) The PEG-model of seasonal succession of planktonic events in fresh waters. Arch. Hydrobiol. 106: 433-471. Starkweather, P.L. (1980) Aspects of the feeding behaviour and trophic ecology of supension feeding rotifers. Hydrobiol. 73: 63-72. Sterner, R.W. (1986) Herbivores' direct and indirect effects on algal populations. Science 231: 605-606. Stockner, J.G. and K . G . Porter. (1988) Microbial food webs in freshwater planktonic cosystems. Chap. 5. In: Complex interactions in lake communities. (Ed: Carpenter,S.R.) Springer-Verlag, New York, N Y , 69-83. Urabe, J. and Y . Watanabe (1991) Effect of food concentration on the assimilation and production efficiencies of Daphnia galeata G.O. Sars (Crustacea: Cladocera). Funct. Ecol. 3: 635-641. Vanni, M.J. (1986) Competition in zooplankton communities: suppression of small species by Daphnia pulex. Limnol. Oceanogr. 31: 1039-1056. Webster, K .E . and R.H. Peters. (1978) Some size dependent inhibitions of larger cladoceran filterers in filamentous suspensions. Limnol. Oceanogr. 23: 1238-1245. Weisse, T. (1990) Trophic interactions among heterotrophic microplankton, nanoplankton, and bacteria in Lake Constance. Hydrobiol. 191: 111-122. Weisse, T. (1991) The annual cycle of heterotrophic freshwater nanoflagellates: role of bottom-up versus top-down control. J. Plankton Res. 13: 167-185. 133 Weisse, T. and H . Muller. (1990) Significance of heterotrophic nanoflagellates and ciliates in large lakes: evidence from Lake Constance. Chap. 29. In: Large Lakes: Ecological structure and function. (Eds: Tilzer,M.M. and C. Serruya) Springer-Verlag, New York, 540-555. Wickham, S.A. and J.J. Gilbert. (1991) Relative vulnerabilities of natural rotifer and ciliate communities to cladocerans: laboratory and field experiments. Freshwater Biology 26: 77-86. Wickham, S.A. and J.J. Gilbert. (1993) The comparative importance of comptetion and predation by Daphnia on ciliated protists. Arch. Hydrobiol. 126: 289-313. Winer, B.J., D.R. Brown, and K . M . Michels.(1991) Statistical principles in experimental design. 3rd ed. McGraw-Hill Inc., Toronto. 1057 p. 134 Appendix 1 Table A . Repeated measures A N O V A of natural log-transformed bacterial abundance in Summer enclosures (5 dates), a=.05 Within-subject Effects Sum of Squares df Mean Square F Sig. Date 1.685 4 .421 12.54 <.001 Date * Treatment 2.610 12 .217 6.472 <.001 Error (Date) .941 28 0.03360 Table B . Repeated measures A N O V A of log-transformed bacterial abundance in Summer enclosures (5 dates), a= .05 Between-subject Effects Sum of Squares df Mean Square F Sig. Treatment 1.959 3 .653 32.586 <.001 Error .140 7 0.02004 Table C . Levene's test of equality of error variances: tests the null hypothesis that the error variance of the natural log-transformed bacterial abundance in the Summer enclosures is equal across sampling dates Sample date F df 1 df 2 Sig . August 9 1.093 3 7 .413 August 12 .759 3 7 .552 August 17 .864 3 7 .503 August 23 4.562 3 7 .045 August 24 14.416 3 7 .002 135 Appendix 1 Table D . Repeated measures A N O V A of natural log-transformed bacterial abundance in F a l l enclosures (5 dates). Within-subject effects Sum of Squares df Mean Square F S ig . Date 2.311 4 .578 13.474 <.001 Date * Treatment 3.670 16 .229 5.348 <.001 Error 1.715 40 .04288 Table E . Repeated measures A N O V A of natural log-transformed bacterial abundance in Fa l l enclosures (5 dates). Between-subject Effects Sum of Squares df Mean Square F S ig . Treatment 1.144 4 .286 18.426 <.001 Error .155 10 .01553 Table F . Levene's test of equality of error variances: tests the null hypothesis that the error variance of the natural log-transformed bacterial abundance in the Fa l l enclosures is equal across sampling dates. Sample date F df 1 df 2 Sig. October 18 1.467 4 10 .283 October 25 2.594 4 10 .101 October 30 1.870 4 10 .192 November 2 6.759 4 10 .007 November 4 3.239 4 10 .060 136 Appendix 1 Table G . Tukey multiple comparison test of natural log-transformed rotifer abundance in enclosures. A l l treatments from both experiments were compared using A N O V A ; significance levels are indicated in the table, with significance at oc= .05 given in bold type. Bonferroni Mul t ip le Comparison Summer Bosmina Summer Daphnia Rotifer Summer F a l l Filament F a l l N o Daphnia Rotifer Grazer Daphnia Fa l l N o +F Grazer Summer Daphnia Bosmina Summer Rotifer Summer N o Grazer Fa l l Daphnia Filament Fa l l Rotifer Daphnia+F Fa l l N o Grazer .001 .009 .033 .989 .003 .004 .708 .384 .004 .946 .898 .999 .057 .002 .976 .995 .999 .053 .278 .991 .176 .869 .012 .933 .971 .017 .264 .746 .840 .024 .910 137 Appendix 2 : Replacement of a missing value Data involving repeated measurements of each experimental unit are appropriately evaluated using repeated measures A N O V A (Winer et al 1991). This analysis was performed using the SPSS 7.5 statistics software, and the procedure does not allow for any missing values in the data set. Thus, of the 8 dates for which samples were counted in the Summer experiment, only 5 sample dates had a complete set of samples from all enclosures. Some extra samples were counted for the D A P H N I A , R O T I F E R and B O S M I N A treatments, but those could not be included as the C O P E P O D enclosure samples were not counted on those dates. Additionally, one sample for the R O T I F E R treatment (enclosure 9) taken on August 12 was inadvertently damaged during processing. Any missing value in the data set results in the entire case (enclosure) being dropped from the analysis. Unfortunately, though the deleted data point occurs early in the experiment when no treatment effects are observed, dropping the entire enclosure from the analysis would influence the results seen at the end of the experiment, where significant results were obtained. With this in mind, I decided to replace the missing value with an estimate and thus allow all the remaining measurements for enclosure 9 to be included in the analysis of bacterial abundance.. To generate an estimate to replace the missing data point, all bacterial abundance values for enclosure 9 were used to generate a linear regression of bacteria density in the bag over the course of the experiment; the predicted value for the bacterial abundance in enclosure 9 on August 12 was then used to replace the missing value in the data set. 138 

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.831.1-0074817/manifest

Comment

Related Items