iv EFFECTS OF TREMATODE PARASITES ON HABITAT USE AND DISTURBANCE RESPONSE OF FRESHWATER SNAILS IN THE GENUS ELIMIA Except where reference is made to the work of others, the work described in this dissertation is my own or was done in collaboration with my advisory committee. This dissertation does not include proprietary or classified information. ________________________________________ Abbie Marie Tomba Certificate of Approval: _________________________ _________________________ Christine A. Sundermann Jack W. Feminella, Chair Professor Associate Professor Biological Sciences Biological Sciences _________________________ _________________________ Gary R. Mullen Stephen L. McFarland Professor Dean Entomology & Plant Pathology Graduate School iv EFFECTS OF TREMATODE PARASITES ON HABITAT USE AND DISTURBANCE RESPONSE OF FRESHWATER SNAILS IN THE GENUS ELIMIA Abbie Marie Tomba A Dissertation Submitted to the Graduate Faculty of Auburn University in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy Auburn, Alabama December 15, 2006 iii EFFECTS OF TREMATODE PARASITES ON HABITAT USE AND DISTURBANCE RESPONSE OF FRESHWATER SNAILS IN THE GENUS ELIMIA Abbie Marie Tomba Permission is granted to Auburn University to make copies of this dissertation at its discretion, upon request of individuals or institutions and at their expense. The author reserves all publication rights. ______________________________ Signature of Author ______________________________ Date of Graduation iv DISSERTATION ABSTRACT EFFECTS OF TREMATODE PARASITES ON HABITAT USE AND DISTURBANCE RESPONSE OF FRESHWATER SNAILS IN THE GENUS ELIMIA Abbie Marie Tomba Doctor of Philosophy, December 15, 2006 (M.S. Bowling Green State University, 1999) (B.S. Bowling Green State University, 1997) 87 Typed Pages Directed by Jack W. Feminella Freshwater snails in the genus Elimia are abundant and important grazers in many southeastern USA streams; in addition, Elimia serves as intermediate hosts to many species of parasitic trematodes and is frequently subject to disturbance from floods. Previous work has shown that parasites can increase Elimia feeding and reduce reproduction; however, the prevalence of trematode infections within Elimia as well as the relationships among parasitism, snail habitat use, and response to disturbance are unknown. I quantified Elimia density, trematode prevalence, and physicochemical habitat conditions, both in wet and dry years, within 7 stream sites in Alabama. All snail populations sampled were parasitized by trematodes to some degree (1?26%). Patterns among streams showed that adult snail density was negatively related to prevalence and percent shade, and was positively related to mean current velocity. v Surveys demonstrated that within-streams, parasite prevalence was positively correlated with snail density and stream substrate size; however, it is unknown whether these patterns resulted from parasites altering habitat use by snails, or because infection rates were higher in these habitats for other reasons. I used in situ enclosures with 2 snail density and 2 substrate size treatments to test if substrate size and snail density affected parasite prevalence and growth of Elimia flava. Compared with their contrasting treatment, the higher snail density and larger substrate treatments both showed significantly higher prevalence. In addition, snail growth was significantly lower in the high- (vs. low-) density treatment. There was no effect of parasite prevalence on snail growth. Last, I used artificial streams to investigate how parasitism, snail size, substrate type, and snail behavior influenced the likelihood of snail dislodgment from substrates during high flows. I placed Elimia flava in artificial streams containing tile or gravel substrates, and then exposed snails to progressively increasing flow velocities. Earlier survey results indicated lower snail density and trematode prevalence in high- (vs. low-) flow years. Artificial stream experiments indicated that snails with high parasite loads dislodged faster than snails with low parasite loads or lacking parasites. Given the widespread distribution of trematodes within Elimia populations and the potential for trematodes to influence snail density both within and between stream sites, trematodes appear to have an important influence on population dynamics of their host snails in many southeastern streams. Parasitic trematodes do not appear to alter host behavior as suggested by others, but, because risk of infection is related to substrate type, snails may potentially alter their behavior to minimize risk of parasitism. vi ACKNOWLEDGMENTS First, I thank my advisor Dr. Jack Feminella for his guidance and patience over the last 6 years. I also thank my committee, Dr. Gary Mullen and Dr. Christine Sundermann, and my outside reader, Dr. Dennis DeVries, for their valuable input on this project and insightful comments on this manuscript. Thanks to all my Auburn friends and lab mates (past and present) for their moral support and for helping me in the field (even if I asked them to help install snail cages for the 3rd time). Special thanks to Mike Gangloff and Lynn Siefferman for being my Auburn family. And finally, my love and thanks to my family, Mom, Dad, Cassie, Amanda, DJ, Olivia and Luke for all their love and support over the years, and for teaching me I can do anything I set my mind to, even if it involves snails. vii Style manual of journal used: Chapter 1 ? Oecologia Chapter 2 ? Journal of the North American Benthological Society Chapter 3 ? Journal of the North American Benthological Society References ? Journal of the North American Benthological Society Computer Software Used: Microsoft Word, Microsoft Excel, JUMP viii TABLE OF CONTENTS LIST OF TABLES.......................................................................................................... x LIST OF FIGURES ........................................................................................................ xii I. INTRODUCTION....................................................................................................... 1 II. CHAPTER ONE: ENVIRONMENTAL FACTORS INFLUENCING ABUNDANCE OF SNAILS IN THE GENUS ELIMIA (PLEUROCERIDAE) AND THEIR TREMATODE PARASITES IN ALABAMA STREAMS, USA ................................................................................ 5 A. INTRODUCTION............................................................................................... 5 B. MATERIALS AND METHODS ........................................................................ 7 C. RESULTS............................................................................................................ .13 D. DISCUSSION ..................................................................................................... .21 III. CHAPTER TWO: INFLUENCE OF TREMATODE PARASITES ON THE DISTURBANCE RESPONSE OF STREAM-DWELLING ELIMIA SNAILS.................................................................................................................... .26 A. INTRODUCTION............................................................................................... .26 B. MATERIALS AND METHODS ........................................................................ .28 C. RESULTS............................................................................................................ .35 D. DISCUSSION ..................................................................................................... .43 ix IV. CHAPTER THREE: INFLUENCE OF SUBSTRATE SIZE AND SNAIL DENSITY ON THE PREVALENCE OF TREMATODE PARASITISM AND GROWTH IN THE STREAM SNAIL ELIMIA FLAVA................................ .50 A. INTRODUCTION............................................................................................... .50 B. MATERIALS AND METHODS ........................................................................ .52 C. RESULTS............................................................................................................ .55 D. DISCUSSION ..................................................................................................... .61 V. REFERENCES........................................................................................................... .65 x LIST OF TABLES Table1-1. Stream names, geographic location, species of the snail Elimia, catchment parent geology, and selected stream physicochemical parameters of the 7 study sites. ..............................................................................9 Table 1-2. Mean (?1 SE) microhabitat variables quantified for the snail Elimia and their trematode parasites. ...............................................................................14 Table 1-3. Mean (?1 SE) densities of total, adult, and juvenile snails in the genus Elimia and % of adult snails parasitized by trematodes (Prevalence), by habitat type at each study site...............................................................................16 Table 1-4. Results of multiple regression analyses indicating environmental variables best predicting adult and juvenile density of snails in the genus Elimia, and % of total Elimia parasitized by trematodes (Prevalence) within each study site. ..........................................................................................20 Table 2-2. Total rainfall (cm) from January to August for 2002, 2003 and 30 y average. Rainfall data was taken from the rainfall station nearest each study site. .............................................................................................................31 Table 2-3. ANOVA table showing effects of substrate size, snail size and parasite load on time to dislodgment for the snail Elimia flava.........................................41 Table 3-1. Table 3-1. 1-way ANOVA and means (?1 SE) for habitat variables across treatment groups. .......................................................................................56 xi Table 3-2. ANOVA table showing effects of snail density and substrate type on parasite load. .........................................................................................................57 Table 3-3. ANOVA table showing effects of snail density and substrate size on snail growth. .........................................................................................................58 xii LIST OF FIGURES Figure 1-1. Mean (?1 SE) density of adult Elimia snails plotted against % of snails parasitized by trematodes (A), % shading (B) and current velocity (C) at the sampling point across the 7 study sites. ........................................................18 Figure 1-2. Size class frequency distribution in high-parasitism sites (Choctafaula site a, Hagood, Flannagin) and low-parasitism sites (Choccolocco, Hatchet, Opintlocco, Choctafaula site b).................................................................................19 Figure 2-1. Mean (+1 SE) snail density across study sites for 2002 (dry year) and 2003 (wet year). ........................................................................................................36 Figure 2-2. Size frequency distributionsof Elimia snails for 2002 (dry year) and 2003 (wet year) for the 7 study sites.........................................................................37 Figure 2-3. Mean (+1 SE) snail width at each site in 2002 (dry year) and 2003 (wet year). ...............................................................................................................38 Figure 2-4. The percentage of small snails (6-8mm) and large snails (>8 mm) parasitized in 2002 and 2003. ...................................................................................40 Figure 2-5. Mean time (?1 SE) to dislodgment of Elimia flava on two substrate treatments (gravel and tiles) across three levels of parasitism (no parasites, low [<150 sporocysts], and high [>150 sporocysts]). The Y- axis also shows the 3 current velocities..............................................................................................42 xiii Figure 2-6. Mean (+1 SE) parasite volume of Elimia flava snails showing downstream and upstream movement.......................................................................44 Figure 2-7. Mean (+1 SE) time to dislodgment for snails of Elimia flava that showed net downstream and net upstream movement..............................................45 Figure 3-1. A.? Mean parasite prevalence sand vs. tile substrate treatments. B.? Mean parasite prevalence in low snail density and high snail density treatments.. ................................................................................................................59 Figure 3-2. A.? Mean growth in the low- vs. the high-density snail treatment. B ? Mean growth in the sand vs. tile substrate treatment.............................................60 1 INTRODUCTION The harsh-benign hypothesis (Peckarsky 1983) predicts that a given system?s disturbance regime will drive which forces influence community structure. Systems with relatively low or infrequent disturbance will be structured by biotic factors such as predation and competition, whereas systems with more frequent or more severe disturbance will be structured by abiotic factors such as flow and scour. In streams, disturbances such as floods can vary greatly in frequency and severity from year to year. Because of this variation, both abiotic and biotic forces may alternate in their importance structuring stream communities over time. Therefore, it is necessary to quantify the interplay between biotic and abiotic factors to understand separate and interactive influences of these factors on communities. In many communities, predation is a biotic factor that greatly affects prey habitat use, foraging (Turner et al. 2000), and reproductive success (Fraser and Gilliam 1992). Parasites occupy the same position in a functional food web as predators, and thus can exact similar impacts on their hosts as predators on prey. Digenean trematodes are a widespread group of parasitic flatworms that use snails almost exclusively as first intermediate hosts (Schmidt and Roberts 1996). Trematodes have a complex life cycle involving several host species, with a vertebrate usually serving as the definitive host. Snails become infected when they ingest eggs from the adult worm or by miracidia, which hatch from eggs and then penetrate the snail. Several larval stages generally 2 develop within snails including sporocysts, redia, and cercaria, with the later stage exiting the snail and infecting the next host (Schmidt and Roberts 1996). Freshwater snails are ubiquitous throughout the southeastern United States. The Family Pleuroceridae contains >150 species, many of which are in the genus Elimia and are subject to low predation because of hard noncrushable shells. Snails in this genus can reach densities of 500 snails /m2 (Dazo 1965) and are the primary grazers in many streams, exerting strong effects on algal biomass and composition (Tuchman and Stevenson 1991, see review Feminella and Hawkins 1995). Elimia also may affect benthic communities by limiting food resources in streams, and thus negatively affecting other grazers (Hill 1992). Large infections of trematodes have been documented in 2 populations of Elimia in the southeastern US (Richardson and Scheiring 1994, Lenertz 2002). To understand and ultimately predict the potential impacts of trematodes on their hosts, and thus stream communities, it is important to 1) quantify the prevalence of trematode infection in their hosts, and 2) identify which environmental factors influence trematode infection within consumers. Several studies have documented the presence of larval trematodes in Elima (Richardson and Scheiring 1994, Lenertz 2002), although the prevalence of infection among several populations and species, and the degree to which infection relates to habitat conditions in streams, are unknown. In marine systems, trematode infection is often highly spatially variable (Lafferty et al. 1994, Smith 2001). For example, in mangrove systems infection rate is strongly related to the number of definitive hosts depositing feces (i.e., parasite source) in an area (Smith 2001). However, in streams, the input of infective material may not be as important as physical properties 3 such as stream current velocity and substrate composition, which may affect the dispersal of infective stages (e.g., miracidium) and thus influence their ability to infect potential hosts. Richardson and Scheiring (1994) found that the riffle-dwelling snail Elimia clara showed lower parasite loads than E. cahawbensis, a species that occurred equally in riffles and pools. Richardson and Scheiring (1994) hypothesized that, because of high flow rates, stream riffles may serve as a physical refuge from trematode infection. High spatial heterogeneity of trematode infection within host snails also may result from direct alteration of host snail habitat use by parasites. For example, trematodes may cause host behavioral modifications, inducing hosts to move to habitats that increase the likelihood of parasite transmission to the next host (reviewed by Poulin 1995, see also Levri 1999). Jokela and Lively (1995) found that snails infected with the trematode Microphallus spp. were more prevalent in shallow shore-bank habitat where snails were more accessible to waterfowl, the definitive host. Conversely, Shinagawa et al. (1999) found that snails infected with several species of trematodes selected deep- water habitats where emerging cercaria could more easily swim to the next host, usually a fish. Shinagawa et al. (1999) hypothesized that habitat shifts by snails may not be adaptive for the parasite, but rather represent either side effects of infection or shifts in resource use. An alternative explanation for the spatial heterogeneity of trematode infection is that differences in trematode prevalence result from differential mortality of infected snails. Parasitized marine snails had a higher mortality rate under low dissolved oxygen concentrations than uninfected snails (Sousa and Gleason 1989). Thus, it is possible that in stressful environments parasitized snails will decrease in abundance because they have a higher mortality rate than unparasitized snails. 4 Stream populations, including pleurocerid snails, or communities also can be influenced by abiotic disturbance, often in the form of floods. Scour and substrate disturbance during floods can result in decreased density and species abundance of benthic organisms (McCabe and Gotelli 2000). Holomuzki and Biggs (1999) demonstrated that frequency of floods affects snail distribution and can greatly reduce snail density. Whereas disturbance may release an individual organism from the pressures of competition, predation, and future parasitic infection, it does not release it from the effects of existing parasitic infections. Parasites deplete host resources for their own development, which can lead to reduced growth (Gerard and Theron 1996) and castration (Krist and Lively 1998, Lenertz 2002) of the host snail. Parasitic infection also can increase host mortality in stressful conditions (anoxia, Sousa and Gleason 1989). Therefore, it is possible that parasitism affects vulnerability of snails to flow disturbances. Last, if trematodes exert strong direct or indirect effects on Elimia habitat use and response to disturbance, then trematodes may play a key role in structuring stream communities under a variety of disturbance regimes. 5 CHAPTER ONE: ENVIRONMENTAL FACTORS INFLUENCING ABUNDANCE OF SNAILS IN THE GENUS ELIMIA (PLEUROCERIDAE) AND THEIR TREMATODE PARASITES IN ALABAMA STREAMS, USA INTRODUCTION Freshwater snails are ubiquitous throughout the southeastern United States. In particular, the family Pleuroceridae contains >150 species, most of which are in the genus Elimia. Pleurocerid snails can reach densities of 500 snails m-2 (Dazo1965) and are the primary grazers in many streams, exerting strong effects on algal biomass, productivity, and species composition (Steinman et al. 1987a, b, Tuchman and Stevenson 1991, Feminella and Hawkins 1995). In this context, Elimia and other pleurocerids may influence benthic communities by limiting food resources and thus negatively affecting other stream grazers (Furnish and Hawkins 1987, Hill 1992). Given the abundance and role of pleurocerid snails in stream communities, it is important to understand the degree to which environmental factors, both biotic and abiotic, contribute to variation in snail abundance and distribution, both within and among streams. Predation is an important biotic factor that may affect the abundance and distribution of prey directly by consumption (Flecker 1984, Eggleston et al. 1997) or indirectly by altering prey behavior (i.e., habitat use, foraging; Gotceitas and Colgan 6 1990, McIntosh and Townsend 1994, Turner et al. 2000), and reproductive success (Fraser and Gilliam 1992, Dorn and Mittelbach 2004). Parasitism also may have pronounced effects on host organisms. Unlike predators, parasites typically do not kill their host (Jokela et al. 1999), although, like predators, parasites occupy the same trophic level and consume valuable prey resources (host tissues). Parasites can alter host behavior and distribution (Bakker et al. 1997, Damsgaard et al. 2005, Gourbal et al. 2001) and limit host reproductive success (Brockerhoff 2004, Newey and Thirgood 2004) in a manner similar to the way predators affect prey. Production of trematode larval stages within their snail hosts represents a significant drain on snail resources, with infection impacting both host reproduction and behavior. Trematodes can affect host reproductive success by destroying host gonadal tissue (castration) chemically, where trematodes interfere metabolically with gonadogenesis (Cheng et al. 1973) or mechanically, when trematodes directly consume gonadal tissue (Probst and Kube 1999). Castration has been demonstrated both in Elimia flava, where infected snails showed decreased gonadal development (Lenertz 2002), and in E. livescens, where infected snails were no longer able to reproduce (Krist 2001). Parsite-induced modification of host behavior may occur when trematodes cause host movement to habitats that increase the likelihood of parasite transmission to the next host (Levri 1999, also see review Poulin 1995). Jokela and Lively (1995) found that Potamopyrgus antipodarum infected with Microphallus spp. were more prevalent in shallow shore-bank habitat than uninfected snails, and thus were more accessible to waterfowl, the definitive host. Conversely, Shinagawa et al. (1999) reported that Semiulcospira libertina infected with several trematode species, were more likely than 7 uninfected snails to select deep-water habitats, where emerging cercaria could more easily access their next host, usually a fish. Given their ability to alter both snail reproduction and habitat use, trematodes may have the potential to influence the abundance and microdistribution of Elimia. Like other organisms, pleurocerid snails persist only in environments for which they are physiologically adapted. Thus, physicochemical habitat variables such as substrate type and dissolved oxygen (DO), may be the ultimate abiotic factors influencing snail abundance and microdistribution, whereas biotic factors, such as parasitism, only modify snail distributions within the abiotic templet (Southwood 1977, Townsend et al. 1997). Therefore, studies of the influence of parasitism on snail distribution also must include measurement of abiotic habitat variables that underlie the basic environmental conditions necessary to parasites and their host snails. Large infections of digenetic trematodes have been documented in 2 populations of Elimia in the southeastern US (Richardson and Scheiring 1994, Lenertz 2002), although the prevalence of infection among different populations and species and the degree to which infection is associated with host habitat conditions, are unknown. The objectives of my study were to 1) determine the extent of trematode infection within several populations of Elimia, 2), quantify the relationships between snail and parasite abundance both within and among streams and stream microhabitats, and 3) quantify the degree to which key abiotic factors are associated with snail abundance and trematode parasitism among snail populations. 8 MATERIALS AND METHODS Study Sites I quantified snail density, their trematode parasites, and several physicochemical variables in 7 stream reaches within 6 streams and 3 river drainages within Alabama, USA (Table 1-1). I selected these sites because preliminary data showed a wide range of parasite prevalence, as indicated by percentage of snails parasitized among sites. All sites were in predominantly forested catchments composed mostly of pine and mixed hardwoods, with many sites either in or adjacent to National Forests (NF). Stream parent geology, which can influence Elimia abundance (Huryn et al. 1995), also varied across sites, ranging from primarily alluvium to primarily limestone, phyllite, or schist (Szabo 1988). Flannagin and Hagood creeks (FL and HAG, respectively; Black Warrior Catchment) are limestone streams within the Bankhead NF, with substrates consisting of exposed bedrock as well as sand and cobble. Geologies at Choccolocco and Hatchet creeks (CL and HAT, respectively; Coosa Drainage) are predominantly phyllite, and chert, phyllite, and schist (Szabo 1988), respectively, with substrate composed mostly of exposed bedrock and cobble. Both streams border the Talladega NF. Opintlocco Creek (OPT, Tallapoosa River Catchment) is primarily alluvial sand and gravel. I selected 2 stream sites on Choctafaula Creek (CFa and CFb, near and within Tuskegee NF, respectively; Tallapoosa River Catchment) because of their contrasting geology (schist and greenstone bedrock vs. alluvial respectively; Szabo1988) and differences in prevalence of snail parasitism (26 and 1%, for CFa and CFb, respectively). CFa was 9 9 Table 1-1. Stream names, geographic location, species of the snail Elimia, catchment parent geology, and selected stream physicochemical parameters of for the 7 study sites. Site (abbreviation) Ecoregion Coordinates Catchment Species Mean stream gradient (%) Parent geology NO3-N (mg L-1) PO4 (mg L-1) pH Specific conduc- tance (?S cm-1) Flannagin Creek (FL) Southwestern Appalachians 34?, 20', 19.6" N 87?, 23', 17.1" W Black Warrior E. cahawbensis 0.58 Limestone 0.243 0.824 7.71 156 Hagood Creek (HAG) Southwestern Appalachians 34?, 20', 38.6" N 87?, 24', 8.5" W Black Warrior E. cahawbensis 0.03 Limestone 0.167 0.867 7.78 81 Hatchet Creek (HAT) Piedmont 33?, 7', 47.0" N 86?, 3', 15.8" W Coosa E. modesta 0.56 Schist/ Greenstone 0.142 0.467 7.82 38 Choccolocco Creek (CL) Piedmont 33?, 49', 39.1" N 85?, 34', 56.4" W Coosa E. modesta 1.34 Phyllite 0.086 0 8.34 60 Choctafaula Creek, site a (CFa) Southeastern Plains 32?, 31', 44.7" N 85?, 33', 19.2" W Tallapoosa E. flava 1.29 Schist/ Greenstone 0.498 0.283 7.39 60 Choctafaula Creek, site b (CFb) Southeastern Plains 32?, 28', 3.7" N 85?, 38', 22.6" W Tallapoosa E. flava 0.50 Alluvial 0.374 0.24 6.05 97 Opintlocco Creek (OPT) Southeastern Plains 32?, 24', 49.1" N 85?, 37', 0.1" W Tallapoosa E. flava 0.26 Alluvial 0.396 0.555 7.13 55 10 located on the fall line, a sharp geologic boundary separating upland piedmont from lowland coastal plains streams (Mount 1975), had substrate consisting mostly of exposed bedrock, sand, and cobble; in contrast, CFb was below the fall line and had mostly sand and gravel substrate. Study species Three species of Elimia occurred in the study area (Table 1-1). Elimia cahawbensis occurred in the Black Warrior River sites (FL and HAG), E. modesta in the Coosa River sites (HAT, CL), and E. flava in the Tallapoosa River sites (CFa, CFb, OPT). Snails with in this genus are dioecious (Dazo 1965, Lenertz 2002), have an estimated life span of 3?4 to 10?11 y (Dazo 1965, Huryen et al. 1994, Richardson and Scheiring 1994), and reach sexual maturity at ~1 y or at 6 mm shell width (Lenertz 2002). Digenean trematodes are a widespread group of parasitic flatworms that use snails almost exclusively as their 1st intermediate hosts (Schmidt and Roberts 1996). They have a complex life cycle involving several host species, with a vertebrate usually as the definitive host. Snails become infected after ingesting eggs from adult worms or by miracidia, which hatch from eggs and then penetrate the snail. Several larval stages generally develop within snails including sporocysts, redia, and cercaria, with the latter stage exiting the snail and infecting the next host (Schmidt and Roberts 1996). Three cercaria types occurred in the sampled Elimia populations including cotylomicrocercous cercaria, virgulate cercaria, and monostome cercaria (Schell 1970, 1985). Cotylomicrocercous cercaria (Family Opecoelidae) typically use fishes as definitive hosts (Schell 1970, 1985). 11 Sampling Snails and parasites.?I sampled each site in Summer 2002, the season when prevalence of trematode parasites (hereafter ?parasites?) is highest for Elimia (Lenertz 2002). 2002 was a relatively dry year, with January to August rainfall across sites averaging ~82 cm (6th driest period between Jan.-Aug. in 50 y; National Weather Service, Birmingham, AL 2002). I conducted surveys during the baseflow period at least 1 wk after a large rainfall event, as high flows after heavy rain reduce snail densities (AMT, personal observations). I quantified snail density from 18 to 20 quadrats per site over a 200- to 300-m reach per site using a Surber sampler (quadrat area = 0.093 m2). However, I sampled only 9 quadrats at HAG because of extremely low-flow conditions. I stratified quadrats evenly between fast-flow (riffle or run) and slow-flow (pool) microhabitats (Hawkins et al. 1993), and sampled each microhabitat randomly. I preserved snails in 95% ethanol in the field, transported them to the laboratory, and then measured them for maximum shell width (nearest 0.01 mm). I classified snails 6? and ?6mm in width as juveniles and adults, respectively (Lenertz 2002), and sexed and dissected adults for trematodes. I removed snails from their shells by fracturing the shell with pliers and using forceps to remove shell fragments. I then used a watchmakers? forceps tip to slice through the body wall along the body length, and probed all tissues under a dissecting microscope (25?50?) for the presence of larval parasites. I did not dissect juveniles (<6mm width) because they are rarely parasitized by trematodes (Lenertz 2002). 12 Microhabitat variables.?I quantified several abiotic microhabitat variables previously shown as correlates of abundances of both Elimia (Huryn et al. 1995, Johnson and Brown 1997, Stewart and Garcia 2002) and their parasites (Jokela and Lively 1995, Shinagawa et al. 1999), including dissolved oxygen (DO, YSI Model 55), streamwater pH (Oakton pHtester3, Model 35624-30), electrical conductivity (YSI Model 33 S-C-T meter), NO3-N (N, DX-120 Ion chromatography, Dionex, Sunnyvale, California), and total P concentration (P, molybdate blue method; Murphy and Riley 1962, Watanabe and Olsen 1965), and current velocity (Marsh-McBirney Flo-mate, Model 2000), substrate size/composition, and % riparian shade (spherical densiometer, Lemmon 1957). I measured current velocity, depth, substrate composition, DO and percent shade at each Surber quadrat. I quantified current velocity in the center of each quadrat, ~1 cm from stream bottom, substrate size by measuring the longest axis of 10 randomly selected particles, and mean % shade by positioning the densiometer immediately above the quadrat and recording measurements from 4 aspects (upstream, downstream, right and left bank directions). Statistical analysis I used a combination of regression approaches to quantify snail?parasite relationships with abiotic variables, at 2 spatial scales: 1) among sites within different streams (large spatial scale); and 2) between runs/riffles and pool microhabitats within individual streams (small spatial scale). For the large-scale analysis, I used simple linear regressions to examine relationships between snail density, prevalence of parasitism (as 13 % of adult snails with parasites, hereafter ?prevalence?) and mean microhabitat variables among the 7 sites. For the small-scale analysis, I used multiple regression (maximum adjusted R-square for model selection, Neter et al. 1996) to determine which abiotic and biotic factors (i.e. current, % shade, substrate size, depth, DO, N, P, prevalence) best predicted adult and juvenile snail abundance within each site. Proportional prevalence and shade data were arcsine-transformed before analysis (Zar 1996). Last, I used Student?s t-test to assess differences in snail densities between runs/riffles and pools within streams, and ?2 to assess differences in prevalence between runs/riffles and pools within streams. I used ? = 0.05 for all analyses except where adjusted for multiple comparisons. RESULTS Microhabitat variables Amount of riparian shading varied greatly among sites, ranging from open sunlit reaches (mean = 17% shade in OPT) to almost entirely shaded reaches (91% in FL; Table 2). Within sites, % shade did not differ between pools and runs. Across sites, mean current velocity ranged from to 0.01 m s-1 at HAG to 0.21 m s-1 at CFb. At the time of sampling (7 August 2002), HAG had flow only in shallow pools, so I sampled snails only from this microhabitat. Mean current was higher in runs than in pools for all other sites except FL (Table 1-2). Mean DO concentration ranged from 5.43 mg L-1 (65% saturation) at HAG to 7.29 mg L-1 (85%) at CL. No measurements were taken at CFa because of a meter malfunction, although subsequent measurements revealed that 14 Table 1-2. Mean (?1 SE) microhabitat variables quantified for the snail Elimia and their trematode parasites. Asterisks indicate significant difference between pools and runs of a given study site (1-way ANOVA, *p<0.01 ). Site abbreviations as in Table 1. ND = data not available. Only pools were sampled in HAG. Site Micro- habitat Shade (% canopy cover) Current velocity (m s-1) Dissolved O2 (mg L-1) Depth (m) Substrate (phi) FL All 91 ? 1 0.06 ? 0.02 5.93 ? 0.12 0.10 ? 0.02 -2.73 ? 0.51 Pool 91 ? 1 0.01 ? 0.01 5.91 ? 0.15 0.14 ?0.02 -2.09 ? 0.62 Run 91 ? 1 0.14 ? 0.02 5.95 ? 0.19 0.04 ?0.03 -3.76 ? 0.78 HAG Pool 89 ? 5 0 5.43 ? 0.46 0.12 ? 0.07 -1.43 ?1.70 HAT All 54 ? 4 0.12 ? 0.03 6.15 ? 0.1 0.16 ? 0.02 -5.92 ? 0.46 Pool 47 ? 6 0.00 ? 0.04* 5.88 ? 0.14 0.18 ? 0.03 -4.81 ? 0.67 Run 59 ? 5 0.18 ? 0.03* 6.32 ? 0.11 0.14 ? 0.02 -6.62 ? 0.53 CL All 82 ? 1 0.07 ? 0.01 7.29 ? 0.16 0.14 ? 0.02 -5.27 ? 0.33 Pool 80 ? 1 0.01 ? 0.01* 6.89 ? 0.15 0.16 ? 0.02 -5.14 ? 0.45 Run 84 ? 1 0.15 ? 0.01* 7.79 ?0.17 0.10 ? 0.02 -5.43 ? 0.51 CFa All 80 ? 11 0.06 ? 0.02 ND 0.23 ? 0.04 -4.65 ? 0.53 Pool 80 ? 4 0.01 ? 0.01* ND 0.25 ? 0.03 -4.21 ? 0.70 Run 79 ? 4 0.14 ? 0.02* ND 0.12 ? 0.04 -5.25 ? 0.25 CFb All 34 ? 3 0.21 ? 0.05 5.72 ? 0.12 0.15 ? 0.02 -2.97 ? 0.92 Pool 35 ? 5 0.00 ? 0.06* 5.22 ? 0.17 * 0.17 ? .03 0.091 ? 1.13* Run 34 ? 4 0.32 ? 0.05* 5.99 ? 0.13* 0.13 ? .02 -5.06 ? 0.83* OPT All 17 ? 2 0.10 ? .03 6.82 ? 0.19 0.14 ? 0.02 -2.58 ? 0.66 Pool 19 ? 2 0.00 ? 0.02* 6.18 ? 0.19* 0.18 ? 0.03 -0.08 ? 0.05* Run 17 ? 2 0.18 ? 0.02* 7.40 ? 0.17* 0.11 ? 0.03 -4.83 ? 0.51* 15 differences in DO between CFa and CFb were negligible (AMT, unpublished data). Mean DO was higher in runs than pools for 2 sites (CFb, OPT; Table 1-2). Mean depth ranged from 0.10 m at FL to 0.23 m at CFa (Table 1-2). Mean depth taken within sampling quadrats did not differ significantly between pools and runs (Table 1-2). Mean substrate size ranged from ?2.73 Phi (fine gravel) at HAG to ?5.92 Phi (very coarse gravel) at HAT (Table 1-2). Mean substrate size did not differ between pools and runs except for 2 of the Coastal Plains streams (CFb and OPT), where substrate sizes in runs were higher than pools (Table 1-2). Snail density and parasite prevalence Overall, mean Elimia density among sites varied by 4-fold, ranging from ~100 snails m-2 (in HAG) to 434 (in HAT) (Table 1-3). Juvenile and adult densities were positively correlated at 3 of the 7 sites (CFa r2 = 0.72, p < 0.001; CFb r2 = 0.69, p < 0.001; OPT r2 = 0.25, p < 0.039). All sites and microhabitats had snails that were parasitized by trematodes, although prevalence varied greatly among sites (Table 1-3). Prevalence ranged from 1 to 26%, with CFa and CFb showing the lowest and highest infections, respectively. Prevalence did not differ between runs and pools for any site (Table 1-3). Cotylomicrocercous cercaria were the most common trematode encountered and composed nearly 60% of infections overall; in addition, they were the most common trematode in CFa and HAT, accounting for 95 and 57% of infections, respectively. The virgulate cercaria (Family Lecithodendriidae) use mammals, birds and amphibians as definitive hosts (Schell 1970, 1985). These cercaria composed 23% of infections overall 16 Table 1-3. Mean (?1 SE) densities of total, adult, and juvenile snails in the genus Elimia and % of adult snails parasitized by trematodes (% Prevalence), by microhabitat at each study site. Asterisks indicate significant difference between pools and runs (t-test, *p<0.01). Site abbreviations as in Table 1-1. ND = data not available Snail density (no. m-2) Site Microhabitat Total Adults Juveniles % Prevalence FL All 282 ? 68 56 ? 11 226 ? 69 23.0 Pool 281 ? 90 49 ? 14 232 ? 92 28.0 Run 284 ? 112 66 ? 18 218 ? 115 20.4 HAG All Pool 100 ?12 37 ? 10 63 ? 13 18.0 Run ND ND ND ND HAT All 434 ? 56 249 ? 29 185 ? 43 5.0 Pool 376 ? 91 238 ? 48 138 ? 70 4.5 Run 470 ? 72 256 ? 38 214 ? 56 5.0 CL All 319 ? 49 192 ? 30 127 ? 34 10.0 Pool 277 ? 66 201 ? 41 76 ? 43 10.2 Run 371 ? 74 181 ? 46 190 ? 48 10.6 CFa All 264 ? 50 133 ? 26 131 ? 26 26.0 Pool 262 ? 68 132 ? 35 130 ? 36 28.8 Run 266 ? 80 134 ? 42 132 ? 42 20.8 CFb All 409 ? 86 232 ? 48 173 ? 43 1.0 Pool 700 ? 124* 376 ? 72 324 ? 59* 0.1 Run 246 ? 91* 155 ? 53 91 ? 44* 1.1 OPT All 268 ? 52 235 ? 46 33 ? 12 5.0 Pool 205 ? 76 173 ? 66 33 ? 18 6.9 Run 324 ? 72 291 ? 62 32 ? 17 3.7 17 and were the most common trematode in CL and FL, accounting for 53% and 69% of infections, respectively. Last, monostome cercaria (Family Pronocephalidae) use birds and turtles are primary definitive hosts (Schell 1970, 1985); these cercaria composed only 18% of infections overall, although they were the most common trematode in CFb, HAG, and OPT (50, 60, and 72% of infections, respectively). Relationships among snail density, prevalence, and microhabitat variables Across the 7 sites, % shade, mean current, and prevalence were the best predictors of snail density. Adult density was negatively related to prevalence (r2 = 0.68, p < 0.02; Fig. 1-1A) % shade (r2 = 0.87, p < 0.01; Fig. 1-1B) and was positively related to current velocity (r2 = 0.66, p < 0.027, Fig. 1-1C). In turn, prevalence was positively related to % shade (r2 = 0.65, p < 0.03). Snail size-class distributions varied significantly between high-prevalence sites (FL, HAG, CFa, prevalence >18%; Table 1-3) and low-prevalence sites (CL, HAT, OPT, CFb, prevalence ?10%; Table 1-3; ?2 = 448, p > 0.0001), with higher frequency of small snails (<3 mm width) in high-prevalence sites and higher frequency of large snails (?5 mm width) in low-prevalence sites (Fig. 1-2). Mean snail width at high-prevalence sites (4.61 ? 0.06) was significantly smaller (t = ?25.74, p < 0.0001) than was mean snail width at low-prevalence sites (6.5 ? 0.04). Within individual sites, prevalence was a significant factor in explaining snail density (either juvenile or adult) in 4 of the 5 (80%) significant multiple regression models (Table 1-4); however, unlike among-stream patterns, prevalence and snail density were positively related at the scale of individual sites. Mean depth (for HAT, CFb) was 18 0 50 100 150 200 250 300 0 5 10 15 20 25 30 % parasitized 0 50 100 150 200 250 300 0 20 40 60 80 100 % shade Ad ult s m- 2 0 50 100 150 200 250 300 350 0 0.05 0.1 0.15 0.2 0.25 Flow m s-1 Figure 1-1. Mean (?1 SE) density of adult Elimia snails plotted against percent of snails parasitized by trematodes (A), % shade, polynomial model (B) and current velocity (C) at the sampling point across the 7 study sites. All 3 relationships were significant. A B C 19 0 20 40 60 80 100 120 <2 2 3 4 5 6 7 8 9 >10 Snail size class (width mm) Fr eq ue nc y m -2 High-prevalence sites Low-prevalence sites Figure 1-2. Size class frequency distribution in high-parasitism sites (Choctafaula site a, Hagood, Flannagin) and low-parasitism sites (Choccolocco, Hatchet, Opintlocco, Choctafaula site b) 20 Table 1-4. Results of multiple regression analyses indicating environmental variables best predicting adult and juvenile density of snails in the genus Elimia, and % of total Elimia parasitized by trematodes (Prevalence) within each study site. Site abbreviations as in Table 1. DO = dissolved oxygen. Adj. R2 = adjusted R2. Snail density (no. m-2) Site Adult Juvenile Prevalence CL Adj R2 p FL Prevalence Adj R2 0.426 p 0.002 HAG Current, DO, prevalence Adj R2 0.8469 p 0.0056 HAT Depth, prevalence Adj R2 0.4673 p 0.0035 OPT % shade Adj R2 0.6848 p 0.0001 CFb Depth, substrate Adj R2 0.4227 p 0.0036 Cfa % shade, prevalence Current, substrate Adj R2 0.4964 0.6709 p 0.0016 0.0001 21 also a significant factor in explaining juvenile densities in 2 of the 5 regression models (40%, Table 1-4). Abiotic factors significantly predicted prevalence within only 2 sites (CFa, OPT). For CFa, mean current, and substrate size were significant (R2 = 0.7294, p < 0.0001), whereas for OPT only percent shade was a significant predictor of prevalence (R2 = 0.6848, p < 0.0001; Table 1-4). Environmental variables pH, N, P, and conductivity showed no relationship to snail densities or parasite prevalence within or across streams. DISCUSSION Hill (1992) hypothesized that Elimia populations are not vulnerable to control by predators because of their hard shells, and are instead subject to bottom-up control through food (periphyton) limitation (see also Hill et al. 1995). My correlative data showing a negative relationship between adult Elimia density and % shade support this hypothesis, as more shaded sites, likely with lower periphyton accrual than less-shaded sites(Kelly et al. 2003), also support lower snail densities than less-shaded systems. In laboratory experiments designed to simulate conditions in Oregon streams, high shade levels both decreased algal biomass and reduced growth of the pleurocerid snail Juga relative to low and intermediate shading (Lamberti et al. 1989). However, in forested streamside channels in Tennessee, increased light levels alone were not sufficient to increases algal biomass (Rosemond 1993), possibly because of nutrient limitation (see also Hill and Knight 1987). High levels of streamwater nutrients such as N and P have also been shown to increase both snail and algal growth (Stelzer and Lamberti 2001, 22 2002), although I found no relationships between streamwater N or P concentrations and Elimia density in my study. My data showing an inverse relationship between adult snail density and prevalence of trematode parasites among streams suggest that, similar to predation, parasitism is an important top-down control on snails (Lafferty 1993). Trematodes may decrease host snail populations by increasing mortality (Sorensen and Minchella 1998, de Montaudouin et al. 2003) or by limiting reproduction by castration of both male and female snails (Krist 2001, Lenertz 2002). In my study, snails in sites with high prevalence of parasitism (18 to 26% of adults parasitized in HAG, FL, and CFa) may have been subject to higher mortality and/or lower reproduction than snails in sites with lower prevalence. Parasitized snails have been shown to consume more periphyton and have lower stored glycogen levels than unparasitized snails, thus implying starvation and potentially increased mortality of parasitized snails (Lenertz 2002). This mechanism would explain lower snail density in high-prevalence sites, although it has not been tested experimentally. Parasitism also may influence snail life history. Differences in size-frequency distributions between high- and low-prevalence sites, with lower proportions of large size classes in high-prevalence sites, also support the hypothesis that parasitism increases snail mortality. Parasitism is more common in large-bodied snails (Lenertz 2002), so increased mortality of large snails from trematode parasitism could explain the lower abundance of large snails in high- vs. low-prevalence streams. Parasitism has also been shown to increase snail growth rate, often resulting in gigantism (Minchilla 1985, Probst and Kube 1999). My results did not support this idea because I found a lower percentage 23 of large snails in high- vs. low-prevalence sites. However, Gorbushin and Levakin (1999) suggest that effects of parasitism on snail growth depend on snail life history, with gigantism being more common in relatively short-lived snails. However, there is little effect of growth on longer-lived snails such as Elimia. Parasite prevalence also was an important factor in predicting juvenile and adult snail densities within individual streams, with 4 of the 5 significant within-stream models indicating a positive relationship between prevalence and density. This pattern seems counterintuitive given that trematodes may increase host mortality (Sorensen and Minchella 1998, de Montaudouin et al. 2003) or decrease host reproduction (Krist 2001, Lenertz 2002); however, several potential explanations exist for this pattern. The positive relationship between snail density and prevalence cannot be explained by classic density- dependant population models common in many infectious diseases (Gilbert et al 1994), as such models predict that increased host density increases transmission because of high parasite-host contact. In most trematode life cycles, however, snails can only become infected from eggs or miracidia, and they cannot transmit parasites to each other (Schmidt and Roberts 1996). The positive relationship between snail density and prevalence could be explained for Elimia if infected snails aggregated more than unparasitized snails. Using y-maze choice tests, Boissier et al. (2003) reported that parasitized Biomphalaria glabrata snails aggregated with other parasitized snails about twice as frequently as with uninfected snails, suggesting that aggregations of infected snails may increase parasite transmission to the next host because cercaria release may be more concentrated from a single host source rather than from isolated individuals. If both infected and uninfected Elimia are attracted to infected Elimia, this pattern would result 24 in high-density patches of snails with high prevalence, and thus could explain why positive relationships between density, both for adult and juveniles, and parasite prevalence occurred. There may be other explanations for the high variation in prevalence among sites. For example, in CFa, the site with the highest prevalence, prevalence was only one of several factors that related to snail density (Table 4). In this site, high prevalence was associated with slow-flowing environments with large substrate. One potential explanation for differential parasite prevalence across microhabitats is that parasites manipulate snail behavior to increase the likelihood of transmission to the next host. Previous studies have shown that trematode infections can alter host snail habitat use and cause hosts to aggregate in specific microhabitats such as at greater depths (Shinagawa et al. 1995) or near shore banks (Jokela and Lively 1995). Parasite alteration of host snail behavior to use slow-flowing habitats may increase parasite transmission success either by 1) causing cercarial release in areas where they are more likely to encounter the next host, which is likely a fish, or 2) reducing the likelihood cercaria are swept downstream by the current. An alternative explanation for differential prevalence across microhabitats may reflect differences in the likelihood of infection between microhabitats. In marine systems, trematode infection is often highly spatially variable (Lafferty et al. 1994). Smith (2001) reported for the snail Cerithidea scalariformis that prevalence in mangrove swamps was highest where birds, the definitive host, commonly perched and deposited feces containing trematode eggs. In streams, however, the input of infective material may not be as important as physical properties such as current, that may affect the 25 dispersal and transmission of snail infective stages (i.e. miracidia) and, thus, their ability to contact potential hosts. Richardson and Scheiring (1994) found that the riffle-dwelling Elimia clara showed lower parasite loads than E. cahawbensis, which occurred equally in riffles and pools. They hypothesized that the high flow rate within riffles serves as a physical refuge from trematode infection (Richardson and Scheiring 1994). In my study, it is possible that infection was more prevalent in areas of slower flows with larger substrates in CFa because the infective stages (eggs or miricidia) are more likely to settle and not be displaced downstream. In addition, large substrates may facilitate trematode contact with snails more readily than that of smaller substrate, if infective stages might become buried within interstitial spaces of the stream bed. All populations of Elimia I sampled were parasitized by trematodes, albeit to different degrees. These results, while compelling in documenting the prevalence of host-parasite interactions are only correlative; further experimental work needs to be done to 1) assess the degree to which trematodes can directly influence snail populations within or among streams in a top-down fashion, and 2) determine the specific mechanisms accounting for strong associations between prevalence and environmental conditions. Given the abundance of trematodes within snail populations and the potential for trematodes to influence snail abundance and distribution on small and large spatial scales, it appears likely that trematodes play an important role in population dynamics of their snail hosts in many streams. 26 CHAPTER TWO: INFLUENCE OF TREMATODE PARASITES ON THE DISTURBANCE RESPONSE OF STREAM-DWELLING ELIMIA SNAILS. INTRODUCTION The harsh-benign hypothesis (sensu Peckarsky 1983) predicts that stream communities with little or infrequent disturbance are structured by biotic factors such as predation and competition, whereas communities with frequent or severe disturbance regimes are structured by abiotic factors such as flow and scour. Research in New Zealand streams supported this hypothesis where greater influences of biotic factors in stable vs. more disturbed streams (Death 2000). However, Thompson et al. (2002) found in Australian streams that disturbance did not alter the affect a predatory stonefly had on mayfly prey, which was contrary to the harsh-benign model. . In streams, floods are a common form of abiotic disturbance (see Lake 2000) which vary greatly in frequency and severity over time (Holomuzki and Biggs 1999) and can influence benthic organisms in several ways. High flows alone and coupled with destabilized substrate can result in both displacement and mortality of stream invertebrates (Holomuzki and Biggs 2000), while scour and substrate disturbance can result in reductions in total abundance and density of species (McCabe and Gotelli 2000). Because of the variation in disturbance frequency and severity, both abiotic and biotic forces may be important in structuring 27 stream communities at different times. Therefore, it is necessary to understand how both biotic and abiotic factors interact to affects stream communities Holomuzki and Biggs (1999) demonstrated that high-frequency flooding can alter snail distribution and reduce density. There are both abiotic and biotic factors that may influence how snails respond to disturbance. Substrate size has been shown to influence the effects of disturbance on snails with gravel substrates providing a refuge from dislodgment over larger substrates (Holomuzki and Biggs 1999) and finer substrates (Moore 1964). Moore (1964) also demonstrated that snail size influenced likelihood of dislodgment in high-flow conditions with small snails being the least susceptible to dislodgment. And finally, aquatic snails have been shown to serve as hosts to trematode parasites (Schmidt and Roberts 1996, Lenertz 2002). While disturbance may release an organism from the pressures of competition, predation, and future parasitic infection, it does not release it from the effects of current parasites. Parasites deplete host resources for their own development, which can lead to reduced growth (Gerard and Theron 1996) and/or castration (Krist and Lively 1998, Lenertz 2002) of the host snail. Parasitic infection can also increase host mortality in stressful environmental conditions (Sousa and Gleason 1989). In this context, it is possible that parasitism may affect the magnitude of snail response to high-flow disturbances, such that parasites may increase host snail vulnerability to flow disturbance. In this study, I compared snail and parasite abundance over two years with strongly contrasting levels of disturbance. My objectives were to 1) determine what biotic and abiotic factors influenced stream snail abundance at different levels of flow disturbance, 2) quantify the influence of increases in flow disturbance on trematode and 28 snail abundance and their interactions and 3) experimentally determine the effects of selected biotic factors (snail size and parasitism), and abiotic factors (substrate size) on snail dislodgment during high flow events. MATERIALS AND METHODS Study Sites I quantified snails, their trematode parasites, and several abiotic physicochemical variables in 7 study stream reaches. The reaches were within 6 streams in 3 ecoregions and 3 river drainages within Alabama (Table 1-1). I selected these sites because preliminary data showed a wide range of parasite prevalence among sites. All sites were in predominantly forested catchments composed mostly of pine and mixed hardwoods, with many sites either in or adjacent to National Forests (NF). Stream geology has been shown to be an important determinant of Elimia abundance (Huryn et al. 1995), and it varied among the study sites from alluvial channels to those composed primarily of limestone, phyllite, or schist (Szabo 1988). Flannagin and Hagood creeks (FL and HAG, respectively; Black Warrior River Drainage) are limestone streams within the Bankhead NF, with substrates consisting of exposed bedrock as well as sand and cobble. Geologies at Choccolocco and Hatchet creeks (CL and HAT, respectively; Coosa River Drainage) border the Talladega NF and are predominantly phyllite, and chert, phyllite, and schist (Szabo1988), respectively, with substrate composed mostly of exposed bedrock and cobble. Opintlocco Creek (OPT, Tallapoosa River Drainage) is primarily alluvial sand and gravel. I studied 2 29 stream sites on Choctafaula Creek (CFa and CFb, near and within Tuskegee NF, respectively; Tallapoosa River Drainage) because of their contrasting geomorphology (schist and greenstone bedrock vs. alluvial, respectively; Szabo 1988) and differences in prevalence of snail parasitism (26 and 1%, respectively). CFa was located on the geologic fall line (Mount 1975), and had substrate consisting of exposed bedrock, sand and cobble, whereas CFb, located below the fall line, had substrate composed largely of sand and gravel. Study species Three species of Elimia occurred in the study area (Table 1-1). Elimia cahawbensis occurred in the Black Warrior River sites (FL and HAG), E. modesta in the Coosa River sites (HAT, CL, and E. flava in the Tallapoosa River sites (CFa, CFb, OPT). Snails within this genus are dioecious (Dazo 1965, Lenertz 2002), have an estimated life span of 3-4 to 10-11 y (Dazo 1965, Huryen et al., Richardson and Scheiring 1994), and reach sexual maturity at ~1 y or ~6 mm shell width (Lenertz 2002). Digenean trematodes are a widespread group of parasitic flatworms that use snails almost exclusively as 1st intermediate hosts (Schmidt and Roberts 1996). They have a complex life cycle involving several host species, with a vertebrate usually serving as the definitive host. Snails become infected after they ingest eggs or from larval miracidia, which hatch from eggs and penetrate snail tissue. Several larval stages generally develop within snails including sporocysts, redia, and cercaria, with the later stage exiting the snail and infecting the next host (Schmidt and Roberts 1996). 30 Field sampling I surveyed each site for Elimia (hereafter ?snails?) and their trematode parasites (hereafter ?parasites?) once in summer 2002 and in 2003, the season when prevalence of trematode parasites is highest (Lenertz 2002). 2002 was a relatively dry year, with January to August rainfall across sites averaging ~82 cm (6th driest Jan.-Aug. period in 50 y, Table 2-1, National Weather Service, Birmingham, AL 2002). In contrast, rainfall for the same period in 2003 was ~130 cm (6th wettest Jan.-Aug. in 50 years, National Weather Service Birmingham, AL 2003). I conducted surveys during the base flow period at least 1 wk after a large rainfall event, as high flows resulting from heavy rain temporarily reduce snail densities (personal observations). I quantified snail density from 18 to 20 quadrats over a 200- to 300-m reach per site using a Surber sample quadrat (area = 0.093 m2). Only 9 quadrats were taken at HAG during 2002 because of extremely low flow conditions and limited microhabitat. I stratified Surber quadrats evenly between fast-flow (riffle or run) and slow-flow (pool) microhabitats (Hawkins et al. 1993), and I sampled each microhabitat randomly. I preserved snails in 95% ethanol in the field, transported them to the laboratory, and then measured them for maximum shell width. Snails ?6 mm in width were classified as adults (Lenertz 2002) and sexed and dissected for trematodes. To assay for trematodes, I removed snails from their shells, made an incision through the body wall, and then microscopically (25 ? 50X) probed all tissues with watchmake?s forceps for presence of 31 Table 2-1. Total rainfall (cm) from January to August for 2002, 2003 and 30-y mean. Rainfall data was taken from the rainfall station nearest each study site. Site abbreviations: Hagood (HAG), Flannagin (FL), Choccolocco (CL), Hatchet (HAT), Choctafaula site a, (CFa), Choctafaula site b (CFb) and, Opintlocco (OPT). Study sites Rainfall station 30-y mean 2002 2003 HAG, FL Addison, AL 146.56 118.80 153.64 CL Heflin, AL 102.97 80.09 134.26 HAT Sylacauga, AL 98.60 83.70 139.57 CFa, CFb, OPT Auburn, AL 94.99 75.39 131.27 32 larval parasite stages. I considered snails with shell widths <6mm as juveniles and enumerated but did not dissect them, as Elimia juveniles are rarely parasitized by trematodes (Lenertz 2002). Abiotic factors.?I quantified several abiotic variables shown in previous studies to relate to both Elimia (Huryn et al. 1995, Johnson and Brown 1997, Stewart and Garcia 2002) and parasite abundances (Jokela and Lively 1995, Shinagawa et al. 1999) at each study site, including dissolved oxygen concentration (DO, YSI Model 55), streamwater pH (Oakton pHtester3, Model 35624-30), electrical conductivity (YSI Model 33 S-C-T meter), NO3-N (N, DX-120 Ion chromatography (Dionex, Sunnyvale, California), total P (P, molybdate blue method, Murphy and Riley 1962, Watanabe and Olsen 1965) concentration, current velocity (Marsh-McBirney Flo-mate, Model 2000), substrate size, and % shade (spherical densiometer, Lemmon 1957). I measured current velocity, depth, substrate composition, DO and % shade at each Surber quadrat. I quantified current velocity in the center of each quadrat, ~1 cm above stream bottom, mean substrate size by measuring the longest axis of 10 randomly selected particles, and mean riparian shade using the densiometer positioned immediately above the quadrat and read from 4 different aspects (upstream, downstream, right and left bank directions). In addition, I calculated link magnitude and used this measure as an index of stream size and catchment area (Osborne and Wiley 1992). 33 Substrate size ? flow experiment I conducted an experiment to quantify the effects of parasite load and snail size, current velocity, and substrate size on the time necessary for snails to become dislodged and carried downstream by the current. At the Auburn University, North Auburn Upper Fisheries Research Station, I constructed 2 artificial stream channels (1.5 m x 10 cm) from PVC pipe cut in half longitudinally and placed within metal stream channels. The outflow ends of the PVC channels were covered with mesh screen to catch dislodged snails. Water was fed by gravity from a 208-L head tank. In each of the channels a 35- cm length of channel 25 cm from the outflow was covered with 1 of 2 substrate treatments. One treatment (unglazed ceramic tile) simulated bedrock, whereas the other treatment consisted of gravel (2?6 cm diameter). I chose these 2 substrate sizes because previous work had shown that the rate of dislodgment varied greatly between them (Holomuzki and Biggs 1999). I used Elimia flava collected from CFa as the focal species for this experiment, a snail and site that showed a relatively high level of parasitism (26%). At the start of each trial, I randomly selected 10 adult snails, individually marked them using a paint pen, and then placed them in one of the 2 channels, allowing snails to acclimate in a low-flow environment for 5 min. After the acclimation period, I ran a 15-min trial where at time 0 flow velocity was increased to ~10 cm/s, ~40 cm/s at 5 min and ~90 cm/s at 10 min. These flow velocities are well within the range snails experience during a high-flow event (unpublished data). I recorded snail movement patterns (i.e., moving upstream, downstream or to the waterline) during the trial and also recorded the time (s) each snail 34 became dislodged. After the experiment, I preserved snails in 95% ethanol and brought them into the laboratory, and I measured their width, dissected them for parasites, and quantified parasite volume and parasite abundance (n = 204 snails assayed). For analysis I used 2 snail size categories, small (6-8 mm in width) and large (? 8 mm) and 3 categories of parasite load: no parasites, low (< 150 sporocysts/snail), and high (? 150 sporocysts/snail). Categories were established by separating the groups at median snail width and parasite load, respectively. I quantified parasite abundance by counting number of larval parasites present in each snail. I estimated parasite volume by first measuring the length and width of 10 randomly selected parasites in each snail. I used the mean width and length to calculate the volume of a cylinder and multiplied this volume by parasite abundance. Statistical analysis I used multiple regression to determine what environmental factors best predicted snail density and/or parasite prevalence across the 7 sites. Proportional parasite prevalence and shade data were both arcsine-transformed before analysis (Zar 1996). I used Chi-square to compare snail size frequency distributions between years (Zar 1996). Snail density between years was compared using the Wilcoxon test (Zar 1996), and one- way ANOVA was used to compare mean snail width between years. Data from the substrate size ? flow experiment were analyzed with a 2-way ANOVA . 35 RESULTS Snail density significantly decreased from 2002 (dry year) to 2003 (wet year) in HAG, CFa, CL, CFb and HAT, but not in FL or OPT (Wilcoxon test, p > 0.05; Fig. 2- 1). Using multiple regression, the best single predictor of adult snail density in 2002 was parasite prevalence (R2 = 0.685 p = 0.022) whereas in 2003 link magnitude (R2 = 0.661 p = 0.022) was the pest predictor. In 2003, snail density not only declined from 2002 in most sites, but there also was a significant shift in snail size distribution at 6 of the 7 sites (CFa, CFb, HAT, HAG, FL, and OPT; ?2 > 57, p > 0.001, Fig. 2-2) compared with 2002, and mean snail size also increased at 6 of 7 sites (CFa, CFb, HAT, HAG, FL, and CL; 1- way ANOVA, p > 0.01, Fig. 2-3). Environmental variables pH, N, P, and conductivity showed no relationship to snail densities or parasite prevalence within or across streams. Three cercaria types occurred in the sampled Elimia populations including cotylomicrocercous cercaria, virgulate cercaria, and monostome cercaria (Schell 1970, 1985). Cotylomicrocercous cercaria are found in the Family Opecoelidae (Lenertz 2002), and opecoelids typically use fishes as definitive hosts (Schell 1970, 1985). Cotylomicrocercous cercaria were the most common trematode encountered and comprised nearly 60% of infections overall, and they were the most common trematode in CFa and HAT (95 and 53% of infections, respectively). The virgulate cercaria were identified as belonging to the Family Lecithodendriiadae and use mammals, birds and amphibians as definitive hosts (Schell 1970, 1985). Virgulate cercaria comprised 27% of infections overall, and they were the most common trematode in CL, HAG and FL (50, 50 and 70% of infections, respectively). Finally, Monostome cercaria belong to the Family Pronocephalidae and birds and turtles are the most common definitive hosts 36 0 100 200 300 400 500 600 CFa CFb CL FL HAG HAT OPT Sn ail s/ m2 2002 2003 Figure 2-1. Mean (+1 SE) snail density at the 7 study sites during 2002 (dry year) and 2003 (wet year). Asterisks indicate significant differences between years. * p>0.05, ** p>0.001. Site abbreviations: Choctafaula site a (CFa), Choctafaula site b (CFb), Choccolocco (CL), Flannagin (FL), Hagood (HAG), Hatchet (HAT) and, Opintlocco (OPT). ** ** ** * ** 37 Choccolocco 0 50 100 150 200 250 300 Hatchet 0 50 100 150 200 250 300 Opintlocco Hagood 0 50 100 150 200 250 300 Choctafaula site a Flannigin 0 50 100 150 200 250 300 <2 2 3 4 5 6 7 8 9 >10 Choctafaula site b <2 2 3 4 5 6 7 8 9 >10 Figure 2-2. Size frequency distributions of Elimia snails for 2002 (dry year) and 2003 (wet year) for the 7 study sites. 2002 2003 Size class Fr eq ue ncy 38 2 3 4 5 6 7 8 9 10 CFa CFb CL FL HAG HAT OPT S he ll w id th (m m) 2002 2003 Figure 2-3. Mean (+1 SE) snail width at each site in 2002 (dry year) and 2003 (wet year). Site abbreviations as in Figure 2-1. Asterisks indicate significant differences between years * p > 0.01, **p > 0.001 ** ** * ** ** ** 39 (Schell 1970, 1985). Monostome cercaria composed only 16% of infections overall; however, they were the most common trematode in CFb and OPT (43 and 65% of infections, respectively). Similar to snail density, the percentage of snails parasitized (hereafter ?prevalence?) also decreased from 2002 to 2003 (?2 = 5.05, p > 0.025). However, there was no difference in the percentage of large snails (>8 mm) parasitized between years (?2 = 0.665, p = 0.415), whereas parasitism in small snails (6-8 mm in width) decreased significantly in 2003 relative to 2002 (?2 = 14.829, p > 0.001, Fig. 2-4). Last, increases in rainfall were positively correlated with changes in the density of small snails (r2 = 0.600, p > 0.04) but were not related to changes in large snail density (r2 = 0.116, p < 0.45). This led to the hypothesis that that small parasitized snails were more susceptible to dislodgment than large parasitized snails. Substrate size ? flow experiment There was a significant effect of parasitism on time to dislodgment, with snails with high parasite loads dislodging faster than snails with low or no parasites (Fig. 2-5). There was no effect of snail size (Table 2-3) or interaction between snail size and parasitism (Table 2-3). There also was no effect of substrate size on dislodgment (Table 2-3); however, there was a significant interaction between substrate size and parasite load (Table 2-3), as snails with low parasite loads dislodged faster on tiles than on gravel substrates (Fig. 2-5). 40 0 2 4 6 8 10 12 14 Small snails Large snails % pa ra sit ize d 2002 2003 Figure 2-4. The percentage of small (6-8mm shell width) and large Elimia snails (>8mm) parasitized by larval trematodes in 2002 and 2003. * indicates p>0.001 (n= 2840) * 41 Table 2-3. ANOVA table showing effects of substrate size, snail size and parasite load on time to dislodgment for the snail Elimia flava. * indicates significant model at 0.05 level DF SS MS F p Model 8 1841666 230208 2.9681 0.0035* Substrate 1 134345 1.7321 0.1894 Snail size 1 110304 1.4222 0.2343 Parasite load 2 556350 3.5866 0.0293* Substrate x Parasite load 2 551998 3.5585 0.0301* Size x Parasite load 2 279661 1.8029 0.1671 Error 230 17838773 77560 Total 19680439 42 300 400 500 600 700 800 900 1000 1100 1200 No Parasites Low High Gravel Tile Figure 2-5. Mean time (?1 SE) to dislodgment of Elimia flava on two substrate treatments (gravel and tiles) across three levels of parasitism (no parasites, low [<150 sporocysts], and high [>150 sporocysts]). The Y- axis also shows the 3 current velocities. Pair wise comparisons: between tile and gravel * p > 0.05, across gravel different letters ( i.e. a, b) indicate p > 0.05. Tim e t o d isl od gm en t ( s) ~10 cm/s ~40 cm/s ~90 cm/s * a a b 43 Parasitism also was associated with snail movement patterns within the artificial streams: snails showing net downstream movement over the trial had higher parasite loads than those that moved upstream or showed no net movement (t = -2.594, p < 0.017, Fig. 2-6). Snail behavior also affected dislodgment probability, as individuals moving upstream (t = -4.534, p< 0.0001, Fig. 2-7) or to the waterline (t = -3.642, p< 0.0003) resisted dislodgment by ramping flows longer than snails not showing those behaviors. DISCUSSION In the relatively dry year (2002) with low levels of flood disturbance, prevalence, a biotic factor, was the best predictor of snail abundance. Whereas, in the wet year (2003), link magnitude, an abiotic factor, best predicted snail density. This result supports the predictions of the harsh-benign hypothesis (Peckarsky 1983), which states biotic factors are more influential during periods of low intensity or infrequent disturbance while abiotic factors are more influential in more frequently disturbed systems. The effects of abiotic factors such as geology (Huryn et al. 1995), light and current (Johnson and Brown 1997), and disturbance (Holomuzki and Biggs 1999) on snail populations have been well studied. Effects of biotic factors such as predation (Turner et al. 2000) and competition (Race 1982) are also well known. However, much less is known about how disturbance mediates the influence on biotic and abiotic factors in stream snail populations. Results from my experiment showed that snails with high parasite loads were more susceptible to dislodgment from high current conditions than snails with low 44 0 5 10 15 20 25 Downstream Upstream pa ra sit e v olu me (m m3 ) Figure 2-6. Mean (+1 SE) parasite volume of Elimia flava snails that showed downstream and upstream movement (p < 0.017). Movement 45 300 400 500 600 700 800 900 1000 1100 1200 Downstream Upstream Figure 2-7. Mean (+1 SE) time to dislodgment for snails of Elimia flava that showed net downstream and net upstream movement (p < 0.0001) ~90 cm/s ~40 cm/s ~10 cm/s Tim e t o d isl od gm en t ( s) Net movement 46 parasite loads or no parasites. Parasitized snails may be more susceptible to dislodgment by current for several reasons. The most likely explanation for this pattern is that consumption of host energy by parasites causes hosts to have less ability to withstand the shear force of water at high velocity. Previous work has shown numerous examples of trematode parasites consuming significant host resources, including all reproductive tissue (castration) (Krist et al.1998, Lenertz 2002,). Parasites do not only limit snail reproduction, but they also may reduce levels of stored glycogen in the body tissue (Lenertz 2002). In my study, individuals with high parasite loads may have less stored energy to withstand the current experienced at high flows than snails with low or no parasites. Another explanation is that parasites may alter snail behavior to increase the likelihood of transmission to the next host, a pattern commonly seen in other snail-parasite interactions (Levri 1999, Shinagawa et al. 1999). Data from the ramped flow experiment showed that snails that moved downstream had higher parasite loads than snails that moved upstream, and that snails that moved downstream were more likely to be dislodged than upstream-moving snails. Elimia has been shown both to orient and display net movement upstream over time (Huryn at al.1994). Parasite-induced movement downstream would cause snails to orient with the spire of their shell facing into the current. This position is much less hydrodynamically efficient than an upstream- facing snail, and would generate lift from water underneath the shell, thus increasing the likelihood of dislodgment (Vogel 1994). Movement or displacement of snails downstream could benefit snail parasites by dispersing the parasitic larval stages to areas where fish, the most likely next host, may be more prevalent. However, it is also possible 47 that higher downstream movement shown by parasitized snails is simply a byproduct of infection, with downstream movement being a result of depleted energy reserves and the tendency for snails to orient themselves in the least energetically expensive position, irrespective of the consequence of higher risk of dislodgment. Contrary to Holomuzki and Biggs (1999), I did not find a significant effect of substrate size alone on dislodgment. I did find a significant interaction between parasitism and substrate size, with snails with low parasite loads being dislodged faster on tile than on gravel; however, there was no difference in time to dislodgment in snails with high parasite loads. The roughness of gravel substrate causes eddies and areas of slower flow to form between substrate particles (Davis and Barmuta 1989). Thus, it appears that gravel substrates act as a refuge against dislodgment for snails with low levels of parasitism. In contrast, it appears that snails with high parasite loads are either still subject to dislodgment even in reduced flow spaces in gravel interstices or they may be less mobile because of low energy stores, being dislodged before they can take refuge between substrate particles than less parasitized snails. My artificial stream experiment did show that level of parasitism affects likelihood of snail dislodgment; however, it did not support the hypothesis that small parasitized snails were more susceptible to dislodgment than large parasitized snails, as there was no interaction between size and parasite load in my analysis. I used naturally infected snails in this experiment, so I had a low sample of small highly parasitized snails (n = 4); this limitation may have reduced statistical power to detect a difference in dislodgment time between highly parasitized large vs. small snails. Other factors, 48 however, may explain why small snails were more susceptible to flow disturbance than larger snails. My field surveys showed that frequencies of small snails declined more than those of larger snails from 2002 (wet year) to 2003; this result was surprising because 1) there was no effect of size on dislodgment probability in the artificial stream experiment, and 2) previous work using E. semicarinata has shown that small snails are actually more common in high-flow habitats than larger snails (Johnson and Brown 1997). `Furthermore, Moore (1964) demonstrated in an artificial stream, small snails of two species (Stagnicola palustris and Physa propinqua) withstand higher current velocities than larger snails. In that study, small snails were dislodged at a similar rate as large on sand and clay but larger snails dislodged faster on coarser substrates (basalt rock and pea gravel). These studies examined the effect of current velocity on snail dislodgment, but they did not take into account the effects of sediment movement and scour during a flood, which may influence snail?s susceptibility to dislodgment. Holomuzki and Biggs (1999) found using artificial stream channels, that dislodgment of snails on loose substrates was positively correlated to the amount of substrate that eroded. However, their study species, Potamopyrgus antipodarum, is much smaller and lighter than Elimia. Further, their results from experiments with artificially stabilized vs. natural substrate suggest that dislodgment, especially on smaller gravel substrates, was much higher on loose substrate vs. stabilized substrate. It is unknown how scour and substrate movement influence snails of different sizes. In addition, if snails are entrained in the water column during a disturbance, small snails should be entrained longer and thus be carried greater distances than larger snails, which may also explain why I found reduced numbers of smaller snails 49 in 2003. Holomuzki and Biggs (1999) also demonstrated that once snails were entrained entrained while mortality was low (~ 6%) it varied with substrate size and was highest amongst coarser substrate (cobble). In addition to scour, substrate deposition may also influence a snail?s response to disturbance. Matthaei et al. (1999) found that flood events could deposit as much as 24 cm of substrate in stream patches. It is possible that the differential response to disturbance between size classes could depend on how snails respond to burial. Burial could lead to conditions such as anoxia or starvation, and parasitism has been shown to increase mortality in both conditions (Sousa and Gleason 1989, Jokela et al. 1999). Further work is needed to determine if size or parasitism influences snail survival during burial or the ability to return to the surface. In conclusion, this study supports the harsh-benign hypothesis by demonstrating that biotic factors such as parasitism best predicted snail abundance in normal water years and abiotic factors such as link magnitude best predicted snail abundance in high-flow years. However, this study also showed that biotic factors such as parasitism can mediate the influences of disturbance, with high levels of parasitism making snails more susceptible to dislodgment during floods. 50 CHAPTER THREE: INFLUENCE OF SUBSTRATE SIZE AND SNAIL DENSITY ON THE PREVALENCE OF TREMATODE PARASITISM AND GROWTH IN THE STREAM SNAIL ELIMIA FLAVA INTRODUCTION Parasitic trematodes commonly use snails as intermediate hosts, and production of trematode larval stages within host snails represents a significant drain on snail resources, with infection potentially affecting host behavior (Levri 1999, also see review by Poulin 1995), growth (Ibrahim 2006), and reproduction (Lafferty 1993). Changes in host behavior can result in a shift in microhabitat use with parasitized snails altering both where and when they are active (Levri and Fisher 2000, Jokela and Lively 1995). Prevalence of trematodes in snail populations varies spatially in both lotic and lentic habitats (Richardson and Scheiring 1994, Jokela and Lively 1995, Smith 2001). Disparate prevalence of parasites within snail populations among habitats can be explained by at least 3 different mechanisms. First, snails may have a higher likelihood of infection in certain habitats because of higher parasite abundance in those habitats. For example, in mangrove systems parasite prevalence in Cerithidea scalariformis is strongly related to the number of wading birds (the definitive hosts) depositing feces (i.e., parasite source) in an area (Smith 2001). Second, parasites may cause behavioral shifts 51 to new habitats by host snails, possibly to maximize transmission to the next host, or as a byproduct of infection. For example, Jokela and Lively (1995) found that infections of Microphallus spp. in Potamopyrgus antipodarum were more prevalent in shallow, shoreline habitats, where snails were more accessible to waterfowl, the definitive host, than were uninfected snails. Third, in habitats with high environmental stress, infected snails may show a higher mortality rate than uninfected snails, thus lowering the proportion of infected snails in stressful habitats. For example, Sousa and Gleason (1989) found that parasitized marine snails of Cerithidea californica had a higher mortality rate under low dissolved oxygen conditions than uninfected snails. In addition to altering habitat use, trematodes also may alter growth rates of host snails. The effects of parasitism on growth are highly variable and taxon specific both for hosts and parasites (Gorbushin 1997, Sousa 1983). Parasites may increase snail growth rates, which may cause gigantism in some species (Probst and Kube 1999, Ibrahim 2006), whereas in others they may simply stunt growth or have no effect (Sousa 1983, Gorbushin and Levakin 1999). Few studies have investigated the mechanisms responsible for the spatial distribution of parasites in natural communities (Shinagawa et al. 1999), or the effects parasites have on growth in pleurocerid snails (Krist 2000). Shinagawa et al. (1999) reported that infected Semiulcospira libertine were more likely to select deepwater habitats, where emerging cercaria could more easily access their next host, usually a fish. Previous work on the stream snail Elimia flava has shown that parasite prevalence is highest in areas with larger substrate (i.e., bedrock) and high snail densities (Tomba and Feminella, Chapter 1). However, it is unknown if these differences in distribution 52 occurred from alteration of host snail movement into such habitats, or from other factors such as higher likelihood of infection or increased host mortality in areas where parasites are prevalent. The objectives of my study were to determine if 1) snail density and substrate size influence parasite prevalence, 2) parasite load influences snail growth, and 3) snail density and substrate type influence snail growth. Taken together these objectives allowed determination of whether the association among high parasite prevalence, large substrate, and high snail density results from the parasite altering snail habitat use or an increasing likelihood of infection in this habitat. MATERIALS AND METHODS Study species Elimia flava (Gastropoda: Pleuroceridae) is a stream snail common throughout the Tallapoosa River Drainage in eastern Alabama. Snails within this genus are dioecious (Dazo 1965, Lenertz 2002), have an estimated life span of 3?4 to 10?11 y (Dazo 1965, Huryen et al., Richardson and Scheiring 1994) and reach sexual maturity at ~1 y (Lenertz 2002). Digenean trematodes are a widespread group of parasitic flatworms that use snails almost exclusively as 1st intermediate hosts (Schmidt and Roberts 1996). They have a complex life cycle involving several host species, with a vertebrate usually as the definitive host. Snails become infected when they ingest eggs from adult worms or by miracidia, which hatch from eggs and then penetrate the snail. Several larval stages generally develop within snails including sporocysts, redia, and cercaria, with the latter stage exiting the snail and infecting the next host (Schmidt and Roberts 1996). Elimia 53 flava commonly serves as the 1st intermediate host for parasitic trematodes in the Family Opecoelidae (Lenertz 2002), with opecoelids typically using a fish as the definitive host (Schell 1970, 1985). Experimental design I conducted an in situ caging experiment with 2 snail density treatments (high and low) and 2 substrate treatments (unglazed ceramic tile and sand) to investigate the degree to which snail density and substrate influence snail infection by parasites (hereafter ?prevalence?). Elimia flava (hereafter ?snails?) was collected from a site on Choctafaula Creek within the Tuskegee National Forrest where the prevalence of parasitism was low (~1% of snails infected, Tomba Chapter 1). I measured snails (width, nearest 0.01 mm) and then uniquely labeled them either by affixing a Bee tag? to their shell or by marking them with a paint pen. I used snails that ranged in shell width from 6.5 to 8.5 mm as these snails represent mature adults that were more likely to become parasitized than juvenile snails (<6 mm), but not so large (i.e. <9 mm) that they were already likely to be infected (i.e. large adults are often parasitized; Lenertz 2000). I randomly assigned snails to cages within 4 treatment groups (sand substrate, high snail density; sand substrate, low snail density; bedrock (tile) substrate, high snail density; bedrock substrate, low snail density; 8 cages per group n =32). Cages (20l x 19w x 8d cm) were constructed out of 10-mm-mesh plastic hardware cloth and lined with 2-mm-mesh fiberglass window screening to contain the snails. High-density cages were stocked with 20 snails = 526/m2 and low-density cages received 10 snails (= 263/m2); these densities were within the 54 range of ambient snail densities in Choctafaula Creek (unpubl. data). For the large size substrate treatments, I used unglazed ceramic tiles (15 x 15 cm) to simulate bedrock and placed a tile in the bottom of each cage, which was then anchored in primarily cobble and boulder habitats of the stream. For the small size substrate treatment, I placed cages in sandy habitats installed such that the lower 2 cm of the cage bottom was buried within the stream bed. I installed cages in Choctafaula Creek several km upstream of the collection site where parasitism is considerably higher (26%, Tomba Chapter 1). I arranged cages in a stratified block design over a 200-m reach of stream and secured them with rebar. Each block consisted of 1 replicate from each of the 4 treatments. I measured current velocity (Marsh-McBirney Flo-mate, Model 2000), % shade (spherical densiometer, Lemmon 1957) and depth at each cage, and attempted to minimize variation in habitat variable across treatments. I removed accumulated debris every 2 to 3 d to maintain flowing conditions for a 9-wk period (23 June 2003 ? 25 August 2005). At the end of the experiment, I removed cages, preserved snails in 95% ethanol, and then measured them for maximum shell width (nearest 0.01 mm). To dissect snails for trematodes, I removed snails from their shells, used a watchmaker?s forceps tip to slice through the body wall along the snail?s length, and then probed all tissues under a dissecting microscope (25?50?) for the presence of larval parasites. I quantified parasite load by simply counting the number of larval parasites per snail. I used ANOVA to test if substrate size and/or snail density influenced parasite prevalence and/or snail growth. I also used linear regression to examine the relationship between parasite load and snail growth. 55 RESULTS Mean current velocity, % shade and depth did not differ among treatments (Table 3-1). Ten cages out of 32 were excluded from the analysis because >75% of the snails had escaped during the experiment. Losses were roughly similar across treatments with 3 cages each excluded from the low- and high- density tile treatment and the high-density sand treatment, and 1 cage excluded from the low-density sand treatment. There was a significant effect of both snail density and substrate size on parasite prevalence (Table 3-2). Prevalence was higher in cages containing tile substrates (vs. sand) and higher in cages with high (vs. low) snail densities (Fig. 3-1). However, there was no interaction between substrate size and snail density (Table 3-2). Prevalence of snails at the source population remained at 1.25 % of snails infected (n = 80) after the experiment. The ANOVA model describing the effect of snail density and substrate size on snail growth was significant (F = 10.049, p < 0.0004, Table 3-3). Snails in the high- density treatment grew an average of 0.26 mm over the 9-wk experiment whereas snails in the low-density treatment grew and average of 0.50mm (p < 0.001, Fig. 3-2A). Substrate size did not significantly affect growth (Table 3-3, Fig. 3-2B) nor was there an interaction between snail density and substrate size (p = 0.350). Last, there was no relationship between snail growth and parasite load (r2 = 0.0003, p = 0.706), or between growth and parasite prevalence (r2 = 0.072, p = 0.227). 56 Table 3-1. 1-way ANOVA and means (?1 SE) for habitat variables across treatment groups. Treatment abbreviations: Sand-Low Snail Density (Sand-Low), Sand-High Snail Density (Sand-High), Tile-Low Snail Density (Tile-Low), Tile-High Snail Density (Tile- High). Treatments Sand-Low Sand-High Tile-Low Tile-High F p n 7 5 5 5 Current velocity (m/s) 0.065 ? 0.026 0.046 ? 0.023 0.082 ? 0.027 0.086 ? 0.033 0.40 0.75 Depth (m) 0.259 ? 0.018 0.270 ? 0.04 0.269 ? 0.031 0.227 ? 0.283 0.44 0.73 % shade 89.2 ? 1.1 89.4 ? 1.5 89.9 ? 1.3 86.3 ? 1.6 1.28 0.31 57 Table 3-2. ANOVA table showing effects of snail density and substrate type on parasite load. Source of variation df SS MS F p Model 3 0.011 0.0037 3.614 0.033 Snail density 1 0.005 5.0224 0.038 Substrate size 1 0.005 4.8524 0.041 Snail density x Substrate 1 0.001 0.5618 0.463 Error 18 0.018 0.0011 Total 21 0.029 58 Table 3-3. ANOVA table showing effects of snail density and substrate size on snail growth. Source of variation df SS MS F p Model 3 0.348 0.1159 10.0486 0.0004 Snail density 1 0.307 26.6425 0.0001 Substrate size 1 0.023 2.0214 0.1722 Snail density x Substrate 1 0.011 0.9307 0.3475 Error 18 0.208 0.0115 Total 21 0.555 59 0 1 2 3 4 5 6 7 8 9 Sand-Low Tile-Low Sand-High Tile-High % pa ra sti ze d Figure 3-1. Mean (+ 1SE) parasite prevalence across 4 treatment groups. * p < 0.04 for sand vs. tile treatments and low snail density vs. high snail density treatments. Treatment abbreviations: same as Table 3-1. * 60 Figure 3-2. A.? Mean growth in the low- vs. the high-density snail treatment. B ? Mean growth in the sand vs. tile substrate treatment. Box ends represent 25th and 75th quartiles, middle horizontal lines indicate the mean, and the whiskers indicate the outermost data point within 1.5 x interquartile range. . * p < 0.0.001 Sand Tile B Gr ow th (m m) 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Low High A * Snail Density Substrate 61 DISCUSSION Results from my experiment support the hypothesis that an increased likelihood of infection in certain habitats is a mechanism responsible for the spatial variation in trematode prevalence in Elimia flava. As such, it refutes the hypothesis that parasites alter snail habitat use in this population, as caged snails in areas of large substrate had a higher prevalence of parasitism than snails caged in sandy areas. If parasites did in fact alter snail behavior as shown by others, then there would be no difference in parasite prevalence between the substrate or snail density treatments. Additionally, there was no interaction between density and substrate on prevalence; thus it appears that host density and substrate operate independently on likelihood of infection, with the highest prevalence occurring in areas with high snail density and coarse substrate, and the lowest prevalence of parasitism occurring in sandy substrates with low snail density. Details of the trematode life cycle provide an explanation why snails may be more likely to become parasitized on large (vs. small) sized substrates. In opecoelid trematodes, the definitive host is usually a fish, and the fish host passes trematode eggs into the water with its feces (Schell 1970, 1985). There, the eggs develop and hatch into ciliated miracidia. Snails then may become infected either by penetration of the miracidia or by ingesting trematode eggs (Schell 1975). The flowing nature of streams, which frequently cause shifting of small substrates such as sand, increases the probability that eggs/larvae landing in sandy environments become buried vs. eggs entrained on coarse substrate. Such burial in sand may, in turn, reduce the likelihood that miracidia or eggs come into contact with the host snail. In contrast, eggs on larger substrates could 62 remain be less subject to burial, and more likely to be ingested by the host, or the hatching miracidia could more easily contact the host. Previous work on other parasitized snail, such as Biomphalaria glabrata has demonstrated that parasitized snails are attracted to each other and thus aggregate together, potentially explaining why parasite prevalence is higher in areas of high (vs. low) snail density (Boissier et al. 2003). This explanation is not supported by my results; instead, my data suggest that snails in high-density areas within a stream have an increase likelihood of infection, as snails in the high-density treatments had higher prevalence than snails in low-density treatments. In most trematode life cycles, parasites are not transmitted from snail to snail (Schmidt and Roberts 1996); therefore, the higher prevalence at higher snail density response I observed cannot be explained by classic density-dependent population models common in infectious diseases (Gilbert et al. 1994). Such models predict that increased host density increases transmission because of high parasite-host contact. It is important to note, however, that larval sporocysts of some opecoelid trematodes can contain adult trematodes, which can then produce eggs and infective miracidia while still inside the snail (e.g., Plagioporus sinitsini, Barger and Esch 2000). Such a life cycle could potentially explain the higher prevalence of parasitism at higher snail densities; however, the abundance of this species in my study area is unknown, although none of the snails I dissected showed adult trematodes within sporocysts (pers. observ.). Thus, the most probable explanation for high parasite prevalence in areas of high snail density may be that large densities of snails are more easily detected by the infective stage (miricidia). Miracidia exposed to snail-conditioned water have been shown to alter their swimming 63 behavior and aggregate at the source of the conditioned water (Nollen 1994, Chernin 1970). In addition, miracidia appear to show a stronger aggregative response when exposed to water from large (vs. small) snails (Chernin 1970). Finelli et al (2000) found the blue crab Callinectes sapidus was more successful and efficient at locating a food odor source when the odor was delivered at a higher rate. They also found that the odor release rate of prey items (clams) was exponentially related to prey size. A high-density patch of snails could have a higher release rate of odor cues and thus would be more likely to attract miracidia and become infected, than a low-density patch. Parasite induced changes in growth have been documented for many snail species (Mohamed and Ishack 1979, Gerard and Theron 1997, Krist and Lively 1998, Mouritsen et al. 1999, Krist 2000), although, the overall result differs widely. Some studies reporting gigantism from parasitism (Mohamed and Ishack 1981, Probst, and Kube 1999, Gorbushin and Levakin 1999), where as others reported a decreased growth in infected snails (Gerard and Theron 1997, Gorbushin and Levakin 1999, Mouritsen et al. 1999) or no difference in growth between infected and uninfected snails (Gorbushin and Levakin 1999, Mouritsen et al. 1999). One explanation for these disparate growth responses to parasitism is that increased growth is more common in shorter-lived semelparous, species and there is less of an effect on growth in longer-lived, iteroparous species (Sousa 1983). This idea was supported by Gourbushin and Levakin?s (1999) study that measured the effect of parasitism on growth of three snail species with different life histories. Elimia flava is a relatively long-lived (4-11 y) iteroparous species, so my results that indicating no relationship between parasite load and growth further support this idea. However, at 64 least one study of a congener E. livecens showed increased growth when parasitized (Krist 2000), so this explanation does not appear universal. Parasitism did not affect growth, but at low snail densities growth nearly doubled relative to high snail densities. This indicates that these snails are food limited at ambient densities, a common pattern in small shady streams (Hill et al. 1995). Being food limited, especially in the high density treatment, may provide another explanation for increased parasite prevalence in the high density treatment. Calow (1974) reported that both Ancylus fluviatilis and Planorbis contortus showed an increase in movement when deprived of food. For E. flava increased movement from low food resources could cause snails to encounter infective eggs or larvae more frequently and thus explain why parasite prevalence is higher in high snail density treatments. In conclusion, it appears that parasite alteration of behavior is not responsible for the spatial variation in parasitism within this snail population and that certain habitats, namely high snail density and coarse substrate habitats, represent a greater risk of infection than do areas of low snail density and finer substrates. It is not known if snails can select habitats to minimize their risk of infection. Additionally, my results further support the idea that parasitism has limited effects on the growth of these iteroparous, long-lived snails. 65 REFERENCES Bakker, T. C. M., D. Mazzi, and S. Zala. 1997. Parasite-induced changes in behavior and color make Gammarus pulex more prone to fish predation. Ecology 78:1098- 1104. Barger, M.A. and G.W. Esch, 2000. Plagioporus sinitsini (Digenea: Opecoelidae): A One-Host Life Cycle. Journal of Parasitology 86:150-153. Boissier, J., E.R. Rivera, and H. Mone. 2003. Altered behavior of the snail Biomphalaria glabrata as a result of infection with Schistosoma mansoni. Journal of Parasitology 89:429-433. Brockerhoff, A. M. 2004. Occurrence of the internal parasite Portunion sp. (Isopoda: Entoniscidae) and its effect on reproduction in intertidal crabs (Decapoda: Grapsidae) from New Zealand. Journal of Parasitology 90:1338-1344. Calow, P. 1974. Some observations on locomotory strategies and their metabolic effects in two species of freshwater gastropods, Ancylus fluviatilis M?ll. and Planorbis contortus Linn. Oecologia 16:149-161. Cheng, T.C., J.T. Sullivan, and K.R. Harris. 1973. Parasitic castration of the marine Prosobranch gastropod Nassarius obsoletus by sporocysts of Zoogonus rubellus (Trematoda): Histopathology. Journal of Invertebrate Pathology 21:183-190. Chernin, E. 1970. Behavioral responses of miracidia of Schistosoma mansoni and other trematodes to substances emitted by snails. Journal of Parasitology 56:287-296 66 Damsgaard, J. T., K. N. Mouritsen, and K. T. Jensen. 2005. Surface activity of Corophium volutator: a role for parasites? Journal of Sea Research 54:176-184. Davis, J.A. and L.A. Barmuta. 1989. An ecologically useful classification of mean and near-bed flows in streams and rivers. Freshwater Biology 21:271-282. Dazo, B.C. 1965. The morphology and natural history of Pleurocera acuta and Goniobasis livescens (Gastropods: Certhiacea: Pleuroceridae). Malacologia 3:1- 80. de Montaudouin, X., H. Blanchet, I Kisielewski, C. Desclaux and G. Bachelet. 2003. Digenean trematodes moderately alter Hydrobia ulvuae population size structure. Journal of the Marine Biological Association of the United Kingdom 83:297-305. Death, R.G. 2000. Habitat stability and interspecific associations among invertebrate assemblages in New Zealand streams. New Zealand Journal of Marine and Freshwater Research. 34:501-506. Diamond, J.M. 1982. Stream geomorphology and benthic predictions as determinants of the population dynamics and life history of the snail Juga plicfera. Journal of Freshwater Biology 1:577-588. Dorn, N. J., and G. G. Mittelbach, 2004. Effects of a native crayfish (Orconectes virilis) on the reproductive success and nesting behavior of sunfish (Lepomis spp.). Canadian Journal of Fisheries and Aquatic Sciences 61:2135-2143. Eggleston, D. B., R. N. Lipcius, and J. J. Grover. 1997. Predator and shelter-size effects on coral reef fish and spiny lobster prey. Marine Ecology Progress Series 149:43- 59. 67 Feminella, J.W., and, C.P. Hawkins. 1995. Interactions between stream herbivores and periphyton: a quantitative analysis of past experiments. Journal of the North American Benthological Society 14:465-509. Finelli, C.M., N.D. Pentcheff, R.K. Zimmer, and D.S. Wethey. 2000. Physical constraints on ecological processes: A field fest of odor-mediated foraging. Ecology 81:784- 797. Flecker, A.S. 1984. The effects of predation and detritus on the structure of a stream insect community: a field test. Oecologia 64:300-305. Fraser, D.F. and J.F. Gilliam. 1992. Nonlethal impacts of predator invasion: facultative suppression of growth and reproduction. Ecology 73:959-970. Gerard, C. and A. Theron. 1996. Altered nutrition and assimilation of th snail host (Biomphalaria glabrata) as a consequence of the parasitic spatial constraint (Schistosoma mansoni). Acta Tropica 61:51-55. Gerard, C. and A. Theron. 1997. Age/size- and time-specific effects of Schistosoma mansoni on energy allocation patterns of its snail host Biomphalaria glabrata. Oecologia 112:447-542. Gilbert, G.S., R.B. Foster, and S. P. Hubbell. 1994. Density and distance-to-adult effects of a canker disease of trees in a moist tropical forest. Oecologia 98:100-108. Gorbushin, A.M. 1997. Field evidence of trematode-induced gigantism in Hydrobia spp. (Gastropoda: Prosobranchia). Journal of the Marine Biological Association of the United Kingdom 77:785-800. 68 Gorbushin, A.M. 1997. Field evidence of trematode-induced gigantism in Hydrobia spp. (Gastropoda: Prosobranchia). Journal of the Marine Biological Association of the United Kingdom. 77:785-800. Gorbushin, A.M. and I.A. Levakin. 1999. The effect of trematode parthenitae on the growth of Onoba aculeus, Littorina saxatilis and L. obtusata (Gastropoda: Prosobranchia). Journal of the Marine Biological Association of the United Kingdom. 79:273-279. Gotceitas, V. and P. Colgan. 1990. The effects of prey availability and predation risk on habitat selection by juvenile bluegill sunfish. Copeia 2: 407-417 Gourbal, B. E. F., M. Righi, G. Petit, and C. Gabrion. 2001. Parasite-altered host behavior in the face of a predator: manipulation or not? Parasitology Research 87:186 ? 192. Hawkins, C. P. and J. K. Furnish. 1987. Are snails important competitors in stream ecosystems? Oikos 49:209-220. Hawkins, C. P., P.A. Bisson, M. Bryant, L. Decker, S. V. Gregory, D. A. McCullough, K. Overton, G. Reeves, R. J. Steedman, and M. Young. 1993. A hierarchical approach to classifying habitats in small streams. Fisheries 18:3-11. Hill, W.R., and A. W. Knight. 1987. Nutrient and light limitation of algae in two Northern California streams. Journal of Phycology 24:125-132. Hill, W.R., M.G. Ryon, and E.M.Schilling. 1995. Light limitation in a stream ecosystem: responses by primary producers and consumers. Ecology. 76:1297-1309. Hill, W.R.. 1992. Food limitation and interspecific competition in snail-dominated streams. Canadian Journal of Fisheries and Aquatic Sciences 49:1257-1267. 69 Holomuzki, J.R., and B.J.F. Biggs. 1999. Distributional responses to flow disturbance by a stream-dwelling snail. Oikos 87:36-47. Holomuzki, J.R., and B.J.F. Biggs. 2000. Taxon-specific responses to high-flow disturbance in streams: implications for population persistence. Journal of the North American Benthological Society 19:670-679. Huryn, A.D., A.C. Benke, and G.M. Ward. 1995. Direct and indirect effects of geology on the distribution, biomass, and production of the freshwater snail Elimia. Journal of the North American Benthological Society 14:519-534. Huryn, A.D., J.W. Koebel and A.C. Benke. 1994. Life history and longevity of the pleurocerid snail Elimia: A comparative study of eight populations. Journal of the North American Benthological Society 13:540-556. Ibrahim, M.M. 2006. Energy allocation patterns in Biomphalaria alexandrina snails in response to cadmium exposure and Schistosoma mansoni infection. Experimental Parasitology 112:31?36. Johnson, P.D., and K.M. Brown. 1997. The role of current and light in explaining the habitat distribution of the lotic snail Elimia semicarinata (Say). Journal of the North American Benthological Society 16:545-561. Jokela, J. and C.M. Lively. 1995. Spatial variation in infection by digenetic trematodes in a population of freshwater snails (Potamopyrgus antipodarum). Oecologia 103:509-517. Jokela, J., C.M. Lively, J. Taskinen, and A.D. Peters. 1999. Effect of starvation on parasite-induced mortality in a freshwater snail (Potamopyrgus antipodarum). Oecologia 119:320-325. 70 Kelly, D.J., M.L. Bothwell, and D.W. Schindler. 2003. Effects of solar ultraviolet radiation on stream benthic communities: An intersite comparison. Ecology 84:2724-2740. Krist, A.C. 2000. Effect of the digenean parasite Proterometra macrostoma on host morphology in the freshwater snail Elimia livescens. Journal of Parasitology 86:262-267. Krist, A.C. 2001. Variation in fecundity among populations of snails is predicted by prevalence of castrating parasites. Evolutionary Ecology Research 3:191-197. Krist, A.C., and C.M. Lively. 1998. Experimental exposure of juvenile snails (Potamopyrgus antipodarum) to infection by trematode larvae (Microphallus sp.): Infectivity, fecundity compensation and growth. Oecologia 116:575-582. Lafferty, K.D. 1993. Effects of parasitic castration on growth, reproduction and population dynamics of the marine snail Cerithidea californica. Marine Ecology Progress Series 96:229-237. Lafferty, K.D., D.T. Sammond, and A.M. Kuris. 1994. Analysis of larval trematode communities. Ecology 35:2275-2285. Lake, P.S. 2000. Disturbance, patchiness, and diversity in streams. Journal of the North American Benthological Society 19:573-592. Lamberti, G.A., S.V. Gregory, L.R. Ashkenas, A.D. Steinman, and C.D. McIntire. 1989. Productive capacity of periphyton as a determinant of plant-herbivore interactions in streams. Ecology 70:1840-1856. Lemmon, P.E. 1957. A spherical densiometer for estimating forest overstory density. Forest Science 2:314-320. 71 Lenertz, K.K.. 2002. Biology, grazing, and physiology of a stream snail Elimia flava (Gastropoda: Prosobranchia) naturally infected by a trematode parasite. MSc Thesis. Auburn University, Auburn, Alabama. Levri, E.P. 1999. Parasite-induced change in host behavior of a freshwater snail: Parasitic manipulation or byproduct of infection? Behavioral Ecology 10:234-241. Levri, E.P. and L.M. Fisher. 2000. The effect of a trematode parasite (Microphallus sp.) on the response of the freshwater snail Potamopyrgus antipodarum to light and gravity. Behaviour 137:1141-1151. McCabe, D.J., and N.J. Gotelli. 2000. Effects of disturbance frequency, intensity, and area on assemblages of stream macroinvertebrates. Oecologia 124:270-279. McIntosh, A. R. and C. R. Townsend. 1994. Interpopulation variation in mayfly antipredator tactics: Differential effects of contrasting predatory fish. Ecology 7:2078-2090. Minchella, D.J. 1985. Host life-history variation in response to parasitism. Parasitology 90:205-216. Mohamed, A.M. and M.M. Ishak. 1981. Growth rate and changes in tissue carbohydrates during schistisome infection of the snail Biomphalaria alexandria. Hydrobiologia 76:17-21. Moore, I.J. 1964. Effects of water currents on fresh-water snails Stagnicola palustris and Physa propinqua. Ecology 45:558-564. Mount, R.H. 1975. The reptiles and amphibians of Alabama. Auburn University, Agricultural Experiment Station. Auburn, Alabama, USA. 72 Mouritsen, K.N., A. Gorbushin, and K.T. Jensen. 1999. Influence of trematode infection on in situ growth rates of Littorina littorea. Journal of the Marine Biological Association of the United Kingdom 79:425-430. Murphy, J., and J.P. Riley 1962. A modified single solution method for the determination of phosphate in natural waters. Analytical Chemistry 27:31-36. National Weather Service Birmingham, Alabama. 2002. Cooperative Rainfall Reports for Year 2002. http://www.srh.noaa.gov/bmx/coop/coops_2002.pdf. Cited 30 June 2005 National Weather Service Birmingham, Alabama. 2003. Cooperative Rainfall Reports for Year 2003. http://www.srh.noaa.gov/bmx/coop/coops_2003.pdf. Cited 30 June 2005 Neter, J., M.H. Kutner, C.J. Nachtsheim, and W. Wasserman. 1996. Applied linear statistical models. McGraw-Hill. Boston, Massachusetts, USA. Newey, S. and S. Thirgood. 2004. Parasite-mediated reduction in fecundity of mountain hares. Proceedings of the Royal Society Biological Sciences Series B 271: S413- S415. Nollen, P.M. 1994. The hatching behavior of Echinostoma trivolvis miracidia and their responses to gravity, light and chemicals. International Journal for Parasitology 24:637-642. Osborne, L. L., and M. J. Wiley. 1992. The influence of tributary spatial position on the structure of warm water fish communities. Canadian Journal of Fisheries and Aquatic Sciences 49:671-681. 73 Peckarsky, B.L. 1983. Biotic interactions or abiotic limitations? A model of lotic community structure. Pages 303-323 in T.D. Fontaine and S.M. Bartell (editors). Dynamics of lotic ecosystems. Ann Arbor Science. Poulin, R. 1995. "Adaptive" changes in the behaviour of parasitized animals: A critical review. International Journal for Parasitology 25:1371-1383. Probst, S. and J. Kube. 1999. Histopathological effects of larval trematode infections in mudsnails and their impact on host growth: What causes gigantism in Hydrobia ventrosa (Gastropoda: Prosobranchia)? Journal of Experimental Marine Biology and Ecology. 238:49-68. Race, M.S., 1982. Competitive displacement and predation between introduced and native mud snails. Oecologia 54:337-347 Richardson, T.D., and J.F. Scheiring. 1994. Ecological observations of two pleurocerid gastropods: Elimia clara (Lea) and E. cahawbensis. Veliger 37:284-289. Rosemond, A.D. 1993. Interactions among irradiance, nutrients, and herbivores constrain a stream algal community. Oecologia 94:585-594. Schell, S.C. 1970. How to know the Trematodes. W.C. Brown. Dubuque, Idaho, USA. Schell, S.C. 1975. The life history of Plagioporus shawi (McIntosh 1939) (Trematoda: Opecoelidae), an intestinal parasite of salmonid fishes. Journal of Parasitology 61:899-905. Schell, S.C. 1985. Handbook of trematodes of North America north of Mexico. University Press of Idaho, Moscow, Idaho, USA. Schmidt, G.D., and L.S. Roberts. 1996. Foundations of parasitology. W. C. Brown Publishers, Dubuque, Idaho, USA 74 Shinagawa, K., M. Urabe, and M. Nagoshi. 1999. Relationship between trematode infection and habitat depth in a freshwater snail, Semiulcospira libertina (Gould). Hydrobiologia 397:171-178. Smith, N.F. 2001. Spatial heterogeneity in recruitment of larval trematodes to snail intermediate hosts. Oecologia 127:115-122. Sousa, W.P. 1983. Host life history and the effect of parasitic castration on growth: A field study of Cerzthzdea calzfornzca Haldeman (Gastropoda: Prosobranchia) and its trematode parasites. Journal of Experimental Marine Biology and Ecology 73:273-296. Sousa, W.P., and M. Gleason. 1989. Does parasitic infection compromise host survival under extreme environmental conditions? The case for Cerithidea californica (Gastropoda: Prosobranchia). Oecologia 80:456-464. Southwood, T.R.E. 1977. Habitat, the templet for ecological strategies? Journal of Animal Ecology 46:336-365. Steinman, A.D., C. D. McIntire, S. V. Gregory, G. A. Lamberti, and L.R. Ashkenas. 1987a. Effects of herbivore type and density on taxonomic structure and physiognomy of algal assemblages in laboratory streams. Journal of the North American Benthological Society 6:175-188. Steinman, A. D., C. D. McIntire, and R. R. Lowry. 1987b. Effects of herbivore type and density on chemical composition of algal assemblages in laboratory streams. Journal of the North American Benthological Society 6:189-197. 75 Stelzer, R.S. and G.A. Lamberti. 2001. Effects of N:P ratio and total nutrient concentration on stream periphyton community structure, biomass, and elemental composition. Limnology and Oceanography 46:356-367. Stelzer, R.S. and G.A. Lamberti. 2002. Ecological stoichiometry in running waters: periphyton chemical composition and snail growth. Ecology 83:1039-1051 Stewart, T.W. and J.E. Garcia. 2002. Environmental factors causing local variation in density and biomass of the snail Leptoxis carinata, in Fishpond Creek, Virginia. American Midland Naturalist 148:172-180. Szabo, M. W. 1988. Geologic map of Alabama [map]. 1:250,000. Geological Survey of Alabama. Tuscaloosa, AL. Thomson, J.R., P.S. Lake, and B.J. Downes. 2002. The effect of hydrological disturbance on the impact of a benthic invertebrate predator. Ecology 83:628-642. Townsend, C.R., S.Doledec, and M.R. Scarsbrook. 1997. Species traits in relation to temporal and spatial heterogeneity in streams: A test of habitat templet theory. Freshwater Biology 37:367-387. Tuchman, N.C., and R.J. Stevenson. 1991. Effects of selective grazing by snails on benthic algal succession. Journal of the North American Benthological Society 10:430-443. Turner, A.M., R.J. Bernot, and C.M. Boes. 2000. Chemical cues modify species interactions: the ecological consequences of predator avoidance by freshwater snails. Oikos 88:148-158. Vogel, S., 1994. Life in moving fluids : the physical biology of flow. Princeton University Press, Princeton, New Jersey. USA. 76 Watanabe, F.S., and S.R. Olsen. 1965. Test of an ascorbic acid method for determining phosphorus in water and NaHCO3 extracts from soil. Soil Science of America Journal 29:677-678. Zar, J.H. 1996. Biostatistical Analysis. 3rd Edition. Prentice Hall. Upper Saddle River, New Jersey. USA.