|USFDC Home | USF Electronic Theses and Dissertations||| RSS|
This item is only available as the following downloads:
Brevetoxins in Marine Birds: Evidence o f Trophic Transfer a nd t he Role o f Prey Fish a s Toxin Vector by Michelle van Deventer A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science College of Marine Science University of South Florida Major Professor: Gabriel A. Vargo Ph.D. Jan H Landsberg, Ph.D. Elizabeth A. Forys, Ph.D. Date of Approval: November 14 2007 Keywords: red tide seabirds, K arenia brevis thread herring, sardines, minnows Copyright 2007 Michelle van Deventer
ACKNOWLEDGEMENTS I would like to thank my advisor, Dr. Gabe Vargo, for his s upport and for giving me the opportunity to be involved in this research. Additional thanks to my committee members, Dr. Jan Landsberg and Dr. Beth Forys, for their insight, guidance and advice during this process. I would like to express additional grat itude to Dr. Gabe Vargo and Dr. Jan Landsberg for their pursuit of funding without which this study would not have been possible. A number of individuals worked on aspects of this project, as well as offering knowledge and advice, including Ms. Karen Atw ood, Ms. Leanne Flewelling and Dr. Danielle Stanek. Additional assistance in identifying fish and stomach contents of birds was provided by staff at the FWCC F ish and Wildlife Research Institute Fisheries Independent Monitoring Lab, which was extremely he lpful. Im grateful for the enthusiastic support I received from local wildlife rehabilitators who provided both birds and anecdotal reports of red tide impacts. Individuals to recognize include Mr. Kevin Barton and Ms. Linda Schrader of the Wildlife Ce nter of Venice, Ms. Barbara Suto at the Suncoast Seabir d Sanctuary in Indian Shores and Dr. P.J. Deitschel, Dr. Amber McNamara and Ms. Robyn Johnson a t the Clinic for the Rehabilitation of Wildlife in Sanibel. Lastly, I owe special thanks to my husband, Paul, and daughter, Meg. They have offered endless support and encouragement, expressed their pride and interest and put up with a lot of birds, some alive and some dead, during the past few years. Funding for this study was provided through a number of sources, including the Florida State Wildlife Grants Program and the Florida Department of Environmental Protection (Award #D0157160). Fish tissue analysis was funded by the Florida Department of Health and Centers for Disease Control through Dr. Jerome Naar at the University or North Carolina Wilmington. Finally, a number of necropsy tools and supplies were purchased with the generous support of the Tampa Bay Parrot Head Club.
i TABLE OF CONTENTS LIST OF TABLES iii LIST OF FIGURES iv ABSTRACT v INTRODUCTION 1 Background and Motivation 1 Harmful Algal Blooms on the West Florida Shelf 2 Brevetoxins and Metabolites 4 Marine Phycotoxins in the Food Web 5 Evidence of HAB Impacts on Marine Birds 8 Overview of the 2005 Karenia brevis Bloom 1 1 Study Obj ectives 14 M ETHODS 15 Study Region 15 Avian Sample Collection 16 Avian Necropsies 16 Fish Sample Collection 17 Tissue Extraction s 18 Brevetoxin Analysis 19 Determination of Bloom Conditions 20 R ESULTS 21 Overview of Avian Samples 21 Brevetoxin Analysis o f Avian Tissues 23 Fish in the Stomachs of Recovered Birds 2 7 Brevetoxin Analysis of Fish Tissues 3 5 D ISCUSSION 41 Brevetoxins in Digestive Tissues of Marine Birds 41 Prey Fish and Brevetoxins in Stomach Contents of Birds 47 Brevetoxins in Common Prey Fish of Marine Birds 50 Management Implications 51 Recommendations for Future Investigations 52
ii REFERENCES 54 APPENDICES 63 Appendix I: Mean weights for all species for which more than one individual was sampled during this study. 6 4 Appendix II: Breve toxin findings for shorebirds collected during reported mortality event in the study region from August to October, 2005. 65 Appendix III: Results for all double crested cormorants ( P. auritus ) collected during this study, and indication of brevetoxicosis diagnosis by atten ding rehabilitation facility. 66
iii LIST OF TABLES Table 1. Mortality rates for sea and shore birds admitted to the Pelican Mans Bird Sanctuary with neurological symptoms during an intense red tide in the fall of 2001. 5 Table 2 Numbers and occurrence of sea birds in Florida whose diet is comprised mainly of small bait fish ( reproduced from Vidal Hernandez and Nesbitt, 2002 with data from Dunning, 1993 ). 12 Table 3 Terms used for Karenia brevis bloom conditions in weekly Red Tide Status reports issued by the FWRI HAB Monitoring Program 20 Table 4 Summary data of bird samples collected during the 2005 Karenia brevis bloom. 2 2 Table 5 Brevetoxin levels in avian tissues collected during from the 2005 Karenia brevis bl oom, as determined by ELISA. 24 Table 6 Brevetoxin levels found in tissue samples from l east t ern chicks recovered between mid May and early July during the 2005 Karenia brevis bloom. 2 7 Table 7 Detailed summary of birds found with whole or partial f ish in stomach during necropsy. 29 Table 8 Distribution and amount of brevetoxin (PbTx) found in the digestive organs and gastrointestinal contents of six marine birds collected from the central west Florida coast during the 2005 red tide event. 33 Table 9 Distribution and amount of brevetoxin (PbTx) found in the digestive organs and gastrointestinal contents of seven marine birds collected from the central west Florida coast during the 2005 red tide event. 3 4 Table 1 0 Distribution of brevetox ins in tissues of baitfish obtained from a local commercial fishery and the food supply of participating rehabilitation facilities during the 2005 K brevis bloom. 37 Table 11 Brevetoxin levels in beached fish collected from beaches along the west cen tral coast of Florida during the 2005 Karenia brevis bloom. 4 0
iv LIST OF FIGURES Figure 1. Maps of the southwest region of Florida (compiled by the Florida Fish and Wildlife Conservation Commission Fish and Wildlife Research Institute) illustrate the development and progression of the 2005 Karenia brevis bloom. 1 3 Figure 2 Map of central west Florida region from Tampa Bay to Charlotte Harbor. 15 Figure 3 The number of birds collected each month for necropsy and brevetoxin analysis during the 2005 K brevis bloom. 21 Figure 4 Comparison of the frequency interval for brevetoxin levels in gall bladder samples of marine birds as compared to liver, stomach and intestinal samples. 2 5 Figure 5. Distribution of brevetoxins in digestive samples f rom cormorants (n=14), brown pelicans (n=6) and other avian species (n=9) for which gall bladders were analyzed. 2 6 Figure 6. Juvenile female d ouble crested cormorant recovered from coastal area of Tampa Bay after exhibiting signs and symptoms of breve toxicosis. 30 Figure 7 A flock of laughing g ulls and r oyal t erns discovered dead on the beach of Siesta Key in Sarasota County, FL on August 25, 2005. 31 Figure 8 Two l aughing g ulls recovered from coastal area of Tampa Bay after exhibiting signs an d symptoms of brevetoxicosis. 32 Figure 9 Brevetoxin levels detected in locally caught thread herring ( O oglinum ) obtained from a local commercial fishery and the food supply of participating rehabilitation facilities. 3 6 Figure 10 Brevetoxin level s detected in mullet ( Mugil spp.), S panish sardines ( S auritus ) and a scaled sardine ( H jaguana ) obtained from a local commercial fishery and the food supply of participating rehabilitation facilities. 37 Figure 11 Brevetoxin levels in small minnow s caught along the west central coast of Florida during the 2005 red tide. 38
v Figure 12 Sanderlings observed actively scavenging red tide killed fish along a beach on the west central coast of Florida during the 2005 Karenia brevis bloom. 39 Figure 1 3 Brevetoxin levels in beached fish collected from beaches along the west central coast of Florida during the 2005 K brevis bloom. 40 Figure 14. Double c rested c ormorants congregated along Casey Key Beach in Sarasota during onshore bloom conditions on March 3, 2005. 43
vi BREVETOXINS IN MARINE BIRDS: EVIDENCE OF TROPHIC TRANSFER AND THE ROLE OF PREY FISH AS TOXIN VECTOR Michelle van Deventer ABSTRACT Harmful algal blooms (HABs) of the brevetoxin producing dinoflagellate Karenia brevis occur periodically along the central west coast of Florida. Mass mortalities of marine birds have long been associated with these blooms, yet there is little data documenting the accumulation of brevetoxins in the tissues of birds and their prey items. An inte nse HAB event impacted the region from Tampa Bay to Charlotte Harbor during most of 2005. More than one hundred marine birds, representing twenty three species, were collected during this bloom All birds sampled were found dead or had died within 24 hou rs of admittance to local wildlife rehabilitation centers I n order t o determine if fish were vectors for brevetoxin ingestion the stomach contents of all birds were examined and any recovered fish were identi fied to the extent possible The gastrointes tinal tissues and contents from all avian samples were analyzed for brevetoxin levels with results ranging from
vii Sh orebirds and gulls may also be exposed to brevetoxins via scavenging of red tide killed fish deposited on beaches du ring blooms. Samples from scavenged fish were found to have brevetoxin levels ranging from 31 to 95,753 ng PbTx per g ram tissue
1 INTRODUCTION Background and Motivation Harmful algal blooms (HABs, or red tides) of the brevetoxin producing dinofla gellate Karenia brevis (previously Gymnodinium breve and Ptychodiscus brevis ) are frequent events in the Gulf of Mexico. There are numerous accounts of avian mortality events during past HABs along the central west coast of Florida (Quick and Henderson, 1 974; Forrester, et al., 1977 ; Kreuder, et al., 2002 ). There is also evidence that catastrophic mortalities of sea birds associated with red tides in Florida have occurred as far back in the historical records as the late Pliocene (Emslie, et al., 1996). Today, rehabilitation centers along Floridas G ulf coast describe treating hundreds of birds exhibiting neurological signs during each month of a red tide. With more than one hundred species of seabirds reported for the Southwest Florida region ( Owre, 199 0 as referenced in Vidal Hernandez and Nesbitt, 2002), a large variety of both migratory and resident birds can be affected when a bloom occurs along the West Florida Shelf (WFS). The precise nature of how K. brevis blooms impact avian populations has no t been thoroughly investigated. Brevetoxins may be a direct cause of death in sea and shorebirds via an acute or lethal exposure or b irds could be exposed to chronic, sublethal levels of toxin over the course of an extended bloom. Subacute doses may als o contribute to morbidity and mortality due to an impaired ability to forage productively, disrupted migration behavior, reduced nesting success, or increased vulnerability to predation, dehydration, disease or injury (Shumway et al., 2003). Past resear ch into phycotoxins and their cycling in the marine food web has revealed that seabirds and other top level marine predators can be exposed to toxins that have bioaccumulated in prey items (Anderson and White, 1992; Tester, et al., 2000; Flewelling, et al. 2005). The brevetoxins produced by K. brevis have been positively identified through tissue analysis as the causative agent in double crested c ormorant ( Phalacrocorax auritus ) mortalities along Floridas Gulf coast (Krueder, et al., 2002),
2 and Vargo et al. (2006) detected low levels of brevetoxins in avian tissues collected during beached bird surveys in Pinellas County during a non bloom period. However, the routes of exposure and specific sources of brevetoxins in avian diets are still unconfirmed and remain poorly understood. The aim of this study was to record the range of brevetoxins present in the gastrointestinal tissues of piscivorous marine birds during a red tide, and to determine whether phytophagous fish can act as vectors of brevetoxins to s ea and shore birds during blooms of K. brevis Piscivorous birds represent apex predators in the marine food web, and understanding sources of brevetoxins in their diet, and the range of brevetoxin levels they might be exposed to, offers insight into pote ntial risks to them and other predatory marine organisms. During this study, s amples of common prey fish, including but not limited to t hread h erring ( Opisthonema oglinum ), S panish s ardines ( Sardinella aurita commonly referred to as whitebait) and anch ovies ( Anchoa spp., commonly referred to as glass minnows), were sampled from various sources during a K. brevis bloom along the central west coast of Florida. Additionally, the stomach contents of marine birds recovered from this region during a bloom were examined to determine recent foraging activity and identify prey items as possible. It is conceivable that avian predators may avoid contaminated prey items during HABs, and so confirmation of brevetoxin levels in gut contents is an important aspect in determining actual exposure via food items. All gastrointestinal tissues, including liver, stomach and/or stomach contents, intestines and/or intestinal contents, and gallbladder, of collected birds were analyzed for brevetoxin levels using a competit ive ELISA (Naar, et al., 2002) Finally, the presence of parent molecules and metabolites were confirmed in representative avian samples with LC MS analysis. Harmful Algal Blooms on the West Florida Shelf Localized K. brevis blooms (>5 x 10 3 cell/l) of relatively short duration occur almost annually along the WFS (Steidinger and Ingle, 1972; Vargo, 1999 ; Tester, et al., 2000). More extensive, longer term blooms are experienced an average of every three to five years (Steidinger and Ingle, 1972). These larger blooms generally last for two to four months, but can range in duration from a few weeks to as long as 18 months.
3 Blooms typically initiate offshore during the late summer or early fall, with August to November being the most common period of onset (Steidinger and Haddad, 1981 ; Vargo, et al., 1987 ). HABs precede the urban development of Floridas Gulf coast, with documented reports dating back to the mid 16th century (Steidinger and Ingle 19 72 ; Turgeon et al., 1998; Vargo, 1999; Kirkpatrick, et a l., 2004). While not restricted to the WFS, the region from Tampa Bay to Charlotte Harbor is where blooms of K. brevis occur most frequently (Steidinger et al., 1998). However, blooms have also occurred in the northern Gulf of Mexico along the coasts of Alabama, Mississippi and Louisiana. There have been occasional blooms along the North Carolina coast as well, as cells are carried by the Loop Current out of the Gulf of Mexico and along the east coast by the Gulf Stream (Anderson and White, 1989; Turgeon et al., 1998). In addition, fish kills along the Delaware coast from July through September 2000 were attributed to a bloom of the brevetoxin producing organism Chattonella cf. verruculosa the first confirmed report of brevetoxin related event in what w as previously thought to be an unaffected area (Bourdelais et al., 2002). Karenia brevis associated red tides along the WFS are consistent with blooms of other toxin producing marine phytoplankton in that they often result in massive fish kills and have be en implicated in the mortalities of marine wildlife such as dolphins (Anderson and White, 1989; Flewelling, et al., 2005), manatees (Bossart et al., 1998) and birds (Kreuder, et al., 2002; Shumway, et al., 2003 ). Research has shown that cell counts of 1 2.5x10 5 cells/L can cause fish kills, and brevetoxicosis can occur at >5x10 3 cells/L (Vargo, 1999). The general area involved in a red tide can be as large as 30,000 km 2 (Steidinger and Ingle, 1972; Turgeon et al., 1998; Vargo, 1999), but K. brevis bloom s are considered to be extremely patchy in nature (Vargo et al., 1987). HABs on Floridas central west coast typically initiate 10 40 miles offshore (Steidinger and Haddad, 1981; Anderson and White, 1989) and records indicate that fish kills can persi st nearly twice as long at offshore areas as compared to inshore areas (Smith, 1975).
4 Brevetoxins and Metabolites The brevetoxins produced by K. brevis are potent lipophilic polyether neurotoxins which bind to voltage gated sodium channels of nerve and muscle cells, causing the cells to become hyperexcitable (Nakanishi, 1985; Colman and Ramsdell, 2003; Baden, et al., 2005). Investigations over the past thirty years have identified two parent toxins, PbTx 1 and PbTx 2, and at least thirteen derivatives w hich are produced by alterations to the reactive side chain of the parent toxins (Baden, et al 2005). Karenia brevis cells produce PbTx 1, PbTx 2 and PbTx 3 in ratios that vary from the growth to dissipation phase, but PbTx 2 is the dominant toxin produ ced at all stages of a bloom (Poli, et al., 2000). PbTx 1 is the most potent brevetoxin and PbTx 3, a reduced form of PbTx 2, is up to ten times more toxic than PbTx 2 when ingested (Baden and Mende, 1982; Landsberg, 2002; Radwan, et al., 2005). Brevetox in derivatives are produced during metabolism and decomposition, and the toxicity of brevetoxin containing food and aerosol is believed to depend on the ratios of parent toxins and metabolites present (Baden, et al., 2005; Poli, et al., 2000; Radwan, et al ., 2005). The profiles of parent toxins and metabolites in an organism may change over time, and it is also likely that the relative quantities can vary by species (Poli, et al., 2000). Investigations into the toxicity of brevetoxins have generally invol ved administration to test subjects via inhalation, intravenous, intraperitoneal or oral routes (Baden and Mende, 1982). Oral exposure is associated with a greater delay in the onset of symptoms versus intravenous or intraperitoneal exposure (5 hours vs. immediate and 30 minutes, respectively; Baden and Mende, 1982). Oral administration of brevetoxins is also associated with a higher LD 50 in mice than intravenous or intraperitoneal administration (0.520 mg/kg vs. 0.094 mg/k g and 0.170 mg/kg, respectively; Baden and Mende, 1982; Fleming and Baden, 1998; ILO, 1984 as referenced in Kirkpatrick, et al., 2004). Toxicokinetic investigations show that tissue uptake, distribution and elimination of brevetoxins varies by exposure route as well (Poli, et al., 1990 ; Cattet and Geraci, 1993; Benson, et al., 2005). In humans, brevetoxin ingestion is responsible for a syndrome referred to as Neurotoxic Shellfish Poisoning, or NSP (Turgeon, et al., 1998; Colman and Ramsdell, 2003; Kirkpatrick, et al., 2004). NSP is cha racterized by both gastrointestinal and
5 neurological symptoms, including nausea, abdominal pain, parasthesia and convulsions (Baden and Mende, 1982; Kirkpatrick, et al., 2004). In sea and shore birds, symptoms of brevetoxin exposure often vary among speci es, but generally include severe ataxia, head tremors, a lack of truncal coordination, lethargy, dehydration and diarrhea (Forrester, et al., 1977; Kreuder, et al., 2002). From September to December 2001, a period of extremely high K. brevis cell concent rations from Tampa Bay to Charlotte Harbor, t he Pelican Man s Bird Sanctuary in Sarasota, Florida treated over 350 birds (including w hite p elicans, b rown p elicans, double crested cormorants, red b reasted m ergansers, g ulls, t erns, skimmers, b oobies, p lovers s andpipers, herons, l oons and ruddy t urnstones) exhibiting symptoms of brevetoxicosis ( Table 1 ). This information is consistent with that from other regions where HABs coincide with increased morbidity and mortality in local bird populations (Hockey and Cooper, 1980; Nisbet, 1983; Work, et al., 1993; Sierra Beltran, et al., 1997; Kreuder, et al., 2002; Shumway, et al., 2003). Table 1. Mortality rates for sea and shore birds admitted to the Pelican Mans Bird Sanctuary with neurological symptoms during an intense red tide in the fall of 2001. Numbers represent summary data from patient log book maintained by PMBS staff. September October November December Total No. of Birds 30 60 107 170 No. of Birds Released 9 31 47 68 No. of Birds Expired 20 29 5 9 102 Mortality Rate 67% 48% 55% 60% Marine Phycotoxins in the Food Web Phytoplankton are the foundation for the oceans food web. Karenia brevis is one of relatively few species of phytoplankton, mostly dinoflagellates, which produce potent toxins (Turgeon, et al., 1998). It is now understood that marine algal toxins, like other pollutants, are capable of biomagnification in the food web with a devastating affect on top level predators (Nisbet, 1983; Anderson and White, 1989; Anderson and White,
6 19 92; Tester, et al., 2000). There is a long history of HABs on the central west coast of Florida, yet our understanding of the vectorial transport of brevetoxins in marine ecosystems is relatively limited. Recent advances in analytical techniques now allo w detection of toxins in much smaller ranges, providing the opportunity for more thorough investigations. This type of analysis is especially valuable in regions such as the WFS, where algal blooms occur with some frequency, may persist for extended perio ds, and the possible exposure pathways for organisms are complex (Tester, et al., 2000). Small, schooling fish play a critical role as food source to predatory fish, mammals and seabirds. They can also provide researchers with a direct link between prim ary producers and top predators. The liver, kidney and other organs of finfish, such as herring, sardine, anchovy, sandlance and mackerel species, can bioaccumulate certain marine toxins and present a risk to organisms that feed on whole fish (Nisbet, 198 3; Anderson, 199 4 ; Turgeon et al., 1998; Lefebvre et al., 1999; Tester, et al., 2000; Lefebvre et al., 2001; Lefebvre et al., 2002b). For example, the northern anchovy ( Engraulis mordax ) has been identified as the primary vector of the Pacific Coast algal toxin, domoic acid, for exposure of pelicans, cormorants, sea lions and other marine animals along the California coast (Lefe bvre, et al., 1999; Scholin, et al., 2000; Lefebvre, et al., 2001; Lefebvre, et al., 2002; Lefebvre, et al., 2002b). In the Atlan tic Ocean, Atlantic mackerel ( Scomber scombrus ) and sand lance ( Ammodytes spp.) have been shown to retain and transfer lethal levels of saxitoxins to predatorial organisms (Nisbet, 1983; Castonguay, et al. 1997). The Gulf of Mexico is critical habitat fo r over 75% of migratory sea and shore birds (Gore 1992) as well as abundant resident species of birds, yet the primary vectors for HAB related morbidities and mortalities have not been identified. It has long been believed that bait fish are killed from t he lethal effects of brevetoxin exposure prior to bioaccumulation at levels that would be dangerous for piscivorous animals (Baden and Mende, 1982; Flewelling et al., 2005). Recent studies indicate this may not always be the case, underscoring the comple x nature of brevetoxins and their role in the regions ecosystem. HABs can be divided into four phases: initiation, growth, maintenance and dissipation (Steidinger and Haddad, 1981; Steidinger and Vargo, 1988; Vargo, 1999). Changes in the frequency, d uration, or intensity of one or more phases during a bloom
7 could influence bioaccumulation of brevetoxins in prey items, and therefore the exposure dynamics for organisms who feed on them. Biomagnification of brevetoxins in fish exposed to sublethal leve ls of K. brevis is an important area of investigation. Studies in the past that focused on single, large doses of toxin known to impair or kill organisms are unlikely to illustrate the actual impacts of brevetoxin cycling in the environment, and their pot ential to contribute to illness and death in avian populations (Kimball and Levin, 1985; Anderson and White, 1989). Recent studies have provided information regarding brevetoxin accumulation in fish exposed to low levels of K. brevis or exposure via prey items. Woofter et al. (2005) monitored the blood of striped mullet ( Mugil cephalus ) exposed to sublethal concentrations of K. brevis in order to determine uptake and elimination rates of brevetoxins for this species In this study, brevetoxin was detect ed fairly rapidly in the blood of exposed mullet, with levels peaking 8 12 hours after initializing exposure and leveling off at approximately 24 hours. Once the mullet were removed from the brevetoxin containing water and placed in control seawater, br evetoxins continued to be detected in the blood for several days. Tester et al. (2000) also verified bioaccumulation of brevetoxins in copepods and subsequent trophic transfer to fish. In this study, fish were not exposed directly to K. brevis in the wat er, but accumulated brevetoxins by feeding upon copepods that had grazed on low concentrations of the dinoflagellate. There is evidence that vario us species and stages of fish have different tolerances to K. brevis bloom conditions, and these tolerance d ifferences may determine whether they are killed by marine algal toxins or bioaccumulate them (Gunter, et al., 1947; Smith, 1975; Steidinger and Haddad, 1981; Tester, et al., 2000). Smaller, benthic species appear to be more susceptible to brevetoxins as they are generally killed earlier and in larger numbers than other faster swimming, pelagic species (Steidinger and Ingle, 1972; Quick and Henderson, 1974; Smith, 1975). Similarly, pelagic fish were found to have higher brevetoxin levels than benthic fish in comparative investigations (Landsberg et al., 2000). There also may be variations in depuration rates depending on the lipid composition or metabolic processes of different fish species (Tester et al., 2000; Colman and Ramsdell, 2003). Lefebvre et al (2002) noted in their study that northern anchovies accumulated
8 domoic acid toxins to a much more potent level than sardines collected simultaneously under identical exposure conditions. Evidence that fish are vectoring brevetoxins to predators in the Gu lf of Mexico is increasing During a stranding event of bottlenose dolphins ( Tursiops truncat u s ) on the eastern Atlantic coast in 1987, the stomach contents of one dolphin, which contained menhaden ( Brevoortia tyran nus ), tested positive for brevetoxins (A nderson and White, 1992). Similarly, in March and April of 2004, more than 100 bottlenose dolphins died along the panhandle region of Florida. M ost of the dolphins stomachs contained fish with high levels of brevetoxins (Flewelling et al., 2005). Addit ional fish collected from the Gulf of Mexico during the 2004 dolphin mortality event also tested positive for elevated levels of brevetoxins, causing investigators to suspect that brevetoxin contaminated prey was an important factor in the stranding of the se mammals (Flewelling et al., 2005). Investigations of brevetoxin accumulation in fish historically focused on accumulation patterns in muscle tissue in order to determine risk to human consumers. Sea and shorebirds, like other marine predators which c onsume whole fish, may be exposed to higher brevetoxin levels in the viscera of fish, as well as cumulatively harmful amounts as they forage over time (Anderson and White, 1992). Unfortunately, sublethal exposures that could lead to disease or death via s econdary effects are difficult to ascertain as these types of ecosystem dynamics are complex and difficult to measure (Quick and Henderson, 1974; Ander son and White, 1992; Tester, et al., 2000). Avian populations, like other wildlife, are vulnerable to nu merous stressors and variables such as disease, pollution, parasitism, competition, habitat loss and predation. It is difficult to determine the impact of an environmental perturbation such as red tide without considering such an event in context of these other parameters. Evidence of HAB Impacts on Marine Birds Phycotoxins have been implicated in morbidity and mortality events for numerous species of birds in various parts of the world (Gunter et al., 1947; Coulson et al., 1968; Forrester et al., 1977; A rmstrong et al., 1978; Nisbet, 1983; Williams et al., 1992; Work et al., 1993; Sierra Beltrn et al., 1997; Krueder et al., 2002; Shumway, et al., 2003).
9 R ecords from the Pelican Mans Bird Sanctuary in Sarasota, Florida show that in the fall of 2001, dur ing an intense red tide in the area, there was an observable increase in the number of avian patients admitted exhibiting neurological signs and symptoms following the September initiation of the bloom. The largest number of bird species admitted occurred in the month of December, nearly three months after the start of the bloom (Table 1). Similarly, a time delay of two to ten weeks was observed by Krueder et al. (2002) between outbreaks of cormorants and K. brevis blooms, with the greatest positive corre lation when bloom levels lagged behind cormorant admittances by eight weeks. Quick and Henderson (1974) reported cormorants, mergansers and lesser scaup dying in red tide affected areas beginning in late February 1974, several months after an October 1973 start of a K. brevis bloom on the WFS. Why marine birds die during red tide blooms in southwest Florida has historically been the subject of debate. There are potentially several ways in which birds can be affected: inhalation of airborne brevetoxins, d rinking or secondarily ingesting K. brevis containing seawater, ingestion of brevetoxins via contaminated prey items, and/or starvation due to reduced food availability. Inhalation of aerosolized brevetoxins are known to cause respiratory irritation in h umans (Pierce, 1986; Kirkpatrick, et al., 2004), lesions of the upper respiratory tract in manatees (Bossart, et al. 1998), and reduced body weight and impaired immune function in rats (Benson, et al. 2005). A great deal of marine bird behavior occurs at the interface of air and water where concentrations of aerosolized brevetoxins occur. It is probable that sea and shore birds are inhaling aerosolized brevetoxins during red tide events, and so it cannot be assumed that this type of exposure is without consequence. However, tissue analysis of cormorants exhibiting signs of brevetoxicosis and collected during bloom conditions showed only very mild pulmonary congestion and no tracheal lesions or significant lung pathology (Krueder et al. 2002). Similar ly, evidence of pulmonary disease was not noted during necropsies of lesser scaup ( Aythya affinis ) found dead or dying during a 1974 red tide in the region (Quick and Henderson, 1974). Benson, et al. (2005) did not observe any neurotoxic effects in rats e xposed to brevetoxin 3 via inhalation for an extended period, and while some physiological effects were observed,
10 no deaths occurred in study subjects. To date, there is no evidence that inhalation of brevetoxins is a primary cause of illness or death of marine birds during red tides. Seabirds possess special adaptations for salt excretion that allows them to drink seawater for hydration, and it is possible that they ingest K. brevis cells in this manner during blooms However, most reports on this beha vior indicate that seabirds are able to manage without relying upon drinking for hydration as they are able t o obtain water directly from their diet, as well as use metabolic water in f at reserves (Elphick, et al., 2001). Also, s eabirds are efficient fe eders. Cormorants and other species will shake and manipulate their prey for easy swallowing, minimizing ingestion of water during foraging (Gunter, 1958; Elphick, et al. 2001). Observations of pelican feeding in the Gulf confirm that this species allow s water to drain from the bill prior to swallowing, thereby minimizing incidental intake of seawater (Gunter, 1958). In the event that some ingestion of toxic water is occurring, Forrester, et al. (1977) exposed domesticated White Pekin ( Anas domesticus ) ducklings to brevetoxin containing seawater and toxic clams in an effort to determine effects. A greater number of ducklings died, and in less time, in the group that received toxic clams and toxic seawater, versus the group that received normal clams and toxic seawater. Close relationships between prey availability and the health of marine bird populations have been confirmed (Anderson, et al. 1982; Schreiber and Cl app, 1987 ). Red tide blooms on the west coast of Florida are associated with short term reduced populations of reef and schooling fish ( Gunter, et al., 1947; Steidinger and Ingle, 1971; Smith, 1975 ). However, necropsies of birds sampled from red tide associated mortalities have foun d individuals with adequate subcutaneous fat reserves and normal breast muscle mass which indicate s an acute cause of death and does not support the theory that all red tide related avian mortalities are a result of starvation (Forrester, et al. 1977 ; Quick and Henderson, 1974 ). The observed delay between initi ation of a bloom and outbreaks in birds has been postulated as the result of the time required for bioaccumulation in schooling fish and eventual transfer to predatory marine birds (Krueder, et al., 2002). In an important discovery that sheds light on pos sible timing and vectorial transport of phycotoxins Flewelling et al. (2005) found that extracellular brevetoxins in seawater, which occur
11 following cell lysis, can accumulate in seagrass and fish and pose a serious threat to marine animals well beyond th e onset of a red tide. While it may not be an obvious route of transfer, prey fish could also pose a risk to shorebirds such as sanderlings ( Calidris alba ), ruddy turnstones ( Arenaria interpres ), plovers ( Charadrius spp.), gulls and other species that fo rage on beaches. It is commonly believed that the primary risk to shorebirds during a red tide is via contamination of shellfish and other invertebrates that comprise their traditional diet. C oquinas ( Donax variabilis ) and other items that shorebirds fee d upon can accumulate marine toxins during red tide blooms and may pose a risk to foraging shorebirds (Cummins et al. 1971; Bretz, et al. 2000). However, this may not be the only route of exposure to brevetoxins for beach dwelling birds. Shorebirds and gulls ( Larus spp.) have been observed actively consuming dead fish when present onshore ( Rand, 1957 ; Gochfeld and Burger, 1980 ). The risk to scavenging shorebirds and gulls when dead fish are deposited on beaches during a red tide event is not well under stood, nor is the distribution of toxins in beached fish. Overview of the 2005 Karenia brevis Bloom A widespread and prolonged bloom of K. brevis impacted the central west coast of Florida for most of 2005. This event began offshore in early January 2005 and was affecting coastal areas from Tampa Bay to Charlotte Harbor within several weeks (Figure 1). The bloom began dissipating in late fall and cell counts in water samples from the region returned to normal in December 2005. This bloom was notable not only for its intensity and duration, but also because bloom conditions combined with warm Gulf waters in late summer result ing in a large area of hypoxia/anoxia and benthic community die offs. The southwest Florida region is home to over 100 species of m arine birds, with b rown p elicans ( Pelicanus occidentalus ), d ouble c rested c ormorants ( P auritus ), terns ( Sterna spp.) and gulls ( Larus spp.) being the most abundant (Table 2). Species which consume marine fish as a primary prey item, or in combination wi th other prey items, include pelicans, herons, egrets, gulls, terns, loons, gannets and cormorants. Most piscivorous marine birds rely on schooling bait fish for food Vidal Hernandez and
12 Nesbitt (2002) estimate that daily consumption rates for these spe ci es range from 135 to 216 g/day, but this range may underestimate actual daily rations for larger seabirds (Bayer, 1989) Table 2. Numbers and occurrence of sea birds in Florida whose diet is comprised mainly of s mall bait fish (reproduced from Vidal He rnandez and Nesbitt, 2002 with data from Dunning, 1993 ). Species Occurrence Population Estimate Brown Pelican Year round 17,000 White Pelican September April 4,000 Double crested Cormorant Year round 20,000 Northern Gannet November May 15,000 Com mon Loon October April 4,000 Terns and G ulls Year round 20,000 Red breasted M erganser s D ucks and G rebes October April 10,000 The very nature of marine birds, predators foraging for prey offshore and then returning to the coast, makes them an ex cellent sentinel species for monitoring ecosystem health, especially when considering the patchy nature of algal blooms in the eastern Gulf of Mexico and the fact that most blooms initiate offshore. Seabirds are perhaps the most conspicuous of marine pred ators, and have often been used as effective monitors of the marine environment (Montevecchi, 1993, 2001). Large numbers of dead or dying seabirds can create an awareness of offshore marine events, and provide important clues of ecosystem disturbances. B irds tend to be very sensitive to marine pollutants and toxins, and have frequently provided the initial evidence for pollution or toxin presence in local waters (Shumway, et al., 2003). Mass mortalities of other marine predators, such as whales, dolphins sharks and large predatory fish, could represent only a fraction of total mortality because most carcasses are likely to drift out to sea or eventually succumb to sinking and scavenging. In contrast, marine birds are capable of flying considerable dista nces from feeding sites where they may have been exposed to toxins in order to
13 return to nesting or roosting grounds where morbidities and mortalities are observed (Kreuder, et al., 2002). Figure 1. Maps of the southwest region of Florida (compiled by the Florida Fish and Wildlife Conservation Commission Fish and Wildlife Research Institute) illustrate the development and progression of the 2005 Karenia brevis bloom. Map A indicates the first water samples offshore containing elevated cell counts durin g the first week of January. Map B shows the presence of the bloom alongshore from Tampa Bay to Charlotte Harbor in late June. Map C, dated mid August, details the extent of the bloom. A B C
14 Study Objectives The objective of this study wa s to identify and quantify the levels of brevetoxins in the gastrointestinal contents and digestive tissues of sea and shore birds during a bloom of K. brevis and to confirm the potential role of prey fish in vectoring brevetoxins to piscivorous marine bi rds. Seabirds often represent apex predators in the marine food web, and their conspicuous nature makes them excellent indicators for the integrity of local fish stocks. Identifying sources of brevetoxins in their diet, and the range of brevetoxin levels they might be exposed to, improves our understanding of how periodic blooms impact avian populations in the region, and offers insight into potential risks for other predatory marine organisms. Floridas birds are an integral part of our wildlife communit ies, and there is evidence that todays populations are only a fraction in size of the colonies that once existed (Runde, 1991 ; Butcher and Niven, 2007 ). Verify ing the trophic transport of brevetoxins to resident or migratory marine bird populations durin g K. brevis blooms is critical for conservation and management planning, particularly when species affected are known to be vulnerable to perturbations in their diet or whose populations are in decline.
15 METHODS Study Region Historically, the regi on from Tampa Bay to Charlotte Harbor (Figure 2) along the central WFS is where K. brevis blooms have occurred most frequently ( Vargo, 1999; Steidinger et al., 1998). During the 2005 bloom, this region experienced massive fish kills as well as reports of marine bird and mammal mortalities. Samples used for this study were gathered from the coastal areas within this region. Figure 2. Map of central west Florida region from Tampa Bay to Charlotte Harbor. Location of rehabilitation facilities and other av ian sample collection sites are indicated.
16 Avian Sample Collection Avian samples used for this study were obtained through a variety of sources. The Seabird Ecological Assessment Network (SEA NET) provided beached birds that were opportunistically recovered during monthly beach surveys on Shell Key in Pinellas County Least tern chicks were provided by the Eckerd College/St. Petersburg Audubon Society Least Tern Rooftop Nest Monitoring project, which monitors the rooftop nes ts of least terns in Pinellas County. All dead least tern chicks observed at nesting sights were collected by observers, labeled, and forwarded to the University of South Florida College of Marine Science (USF CMS) for analysis. T hree independent wildli fe rehabilitation centers in the region (Figure 2) provided carcasses of sea and shore birds that were brought to their facilities but did not survive The Clinic for the Rehabiliation of Wildlife (C.R.O.W.) is located in Sanibel, FL alongside Charlotte H arbor. The Wildlife Center of Venice (W.C.V.) is located in Venice, FL near Sarasota Bay. The Suncoast Seabird Sanctuary (S.S.S.) is located in Indian Shores, FL on Tampa Bay. Combined, these three centers provide wildlife recovery services for nearly 2 00 miles of coastline along the central west coast of Florida and are staffed with individuals experienced in identifying symptoms consistent with suspected brevetoxicosis in seabirds. Only birds which did not eat while in care at a rehabilitation center and expired within 24 hours of intake were used for this study. Records for each bird include location and date found, symptoms presented if alive at time of recovery, and final disposition (e g. died or euthanized) All expired animals were stored froze n in a 20 o C freezer pending necropsy and tissue analysis. Avian Necropsies All avian necropsies were performed in St. Petersburg, FL at USF CMS or at the Florida Fish and Wildlife Conservation Commission Fish and Wildlife Research Institute (FWCC FWRI). Necropsy p rotocols established by SEANET at Tufts Center for Conservation Medicine and the Avian Necropsy Manual for Biologists (Work, 2000) were followed. Notes on carcass and body condition were kept for each bird, including feather condition, muscle mass, subcutaneous fat reserves, stomach contents, and parasite
17 load. At minimum, ventral and dorsal digital photos of each bird were taken. Approximate age was determined, and measurements of tarsus, wing chord, culmen, beak depth and beak width were r ecorded. The gastrointestinal tract was examined for parasites, abnormalities, and evidence of recent feeding. All significant or unusual findings were noted and photographed. All stomach contents were photographed, identified to the extent possible wi th assistance from staff at FWCC FWRI, and analyzed individually for brevetoxins. All tissue samples and stomach contents collected during necropsy were either immediately extracted or placed in individual Whirlpak bags and frozen at 20 o C for future use. Fish Sample Collection Fish samples for this study were taken from several sources during the 2005 bloom of K. brevis Locally caught thread herring ( O oglinum ), S panish sardines ( H jaguana ) and bay anchovies (glass minnows, A mitchilli ) were purch ased monthly from a large commercial fishery in Cortez, FL which supplies many area facilities with locally caught bait fish. These fish were caught in w aters between Sarasota Bay and Tampa Bay but can not be correlated to specific locations or bloom cond itions due to lack of records for catch date and location. No fish were collected after the end of September due to a crash in fish stocks as a reported by local fisherman. The lack of available bait fish was attributed by suppliers to red tide condition s and the associated hypoxic/anoxic zone occurring in the region at the time All fish were stored frozen in a 20 o C freezer pending dissection and tissue analysis. Fish s amples were also taken monthly from the food supply of participating rehabilitation facilities. Rehabbers generally rely on locally caught finger mullet ( M ugil spp. ), thread herring and S panish sardines that have been purchased from regional baitfish companies or commercial suppliers. Once each month, fish were taken from the supply o ffered to patients that day so as to determine the brevetoxin levels present in the food provided injured and sick marine birds treated during a red tide bloom. Samples were stored frozen in a 20 o C freezer pending dissection and tissue analysis. In order to determine the risk to birds foraging on local beaches during a bloom, red tide killed fish deposited onshore during active scavenging by gulls and shorebirds
18 were also collected. Information about date, location and bloom conditions were recorded for all beached fish collected. Finally, additional small prey fish were collected live from coastal areas, with an effort to collect during observed sea and shore bird feeding, or a lack of obvious foraging activity was noted. Date, location and bloom condi tions were recorded for all small minnows collected alongshore and samples were stored frozen in a 20 o C freezer pending dissection and tissue analysis. Tissue Extraction s Sample extractions for avian and fish tissues were performed in St. Petersburg, F L at the USF CMS or the FWCC FWRI Biotoxins Laboratory. Samples of whole tissues, such as liver and gall bladder, were macerated (Waring Commercial Laboratory Blender) and weighed to .1g. The stomachs of larger birds were opened and contents scraped, c ombined, weighed and transferred to 50mL Falcon tubes Any whole fish present in stomach contents were separated and analyzed as described below. Small and large i ntestines were emptied to create a homogenous mixture and a sample removed and weighed for extraction. Minimum sample size requirements for analysis by ELISA required that stomachs and intestines of smaller birds ( species weighing less t han 40 50g) or partially decomposed samples were macerated whole. All samples were weighed and stored in i ndividual 50 mL Falcon tubes for extraction. All fish weighing less than 5g were macerated whole and a representative sample taken for extraction and analysis. For larger fish, all viscera was removed from the intraceolomic cavity and macerated for whole viscera analysis. The exception to this was gravid females, for which eggs were removed, macerated, and sampled separately from visceral organs. A sample of the fish muscle tissue was also removed with skin and scales, macerated and a sample weighed out for analysis. All macerated and weighed tissue samples were extracted twice in 80% methanol at 60C and centrifuged twice (10 minutes each, 3,000 x g ; Eppendorf Centrifuge 5810 ). Both methanol extracts were combined and partitioned with hexanes (1:1 v:v ). Methanolic fractions were transferred to 8 ml Wheaton glass specimen vials with rubber lined caps and stored at 20C until further analysis.
19 Brevetoxin Analysis An enzyme linked immunoabsorbant assay (ELISA) was used to test methanol extract sample s for brevet oxin content (PbTx) All ELISA analysis was performed by staff at the FWCC FWRI Biotoxins Laboratory in St. Petersburg, FL. This is an FDA certified laboratory for brevetoxin analysis and the official state laboratory for shellfish testing du ring HABs. Th e ELISA recognizes all congeners and metabolites of brevetoxin that ha ve a PbTx 2 type backbone and the lower limit of quantification is 2 5 ng/g. P rotocols established by Naar et al. (2002) were followed. To summarize, 96 well microplate s we re pre coasted with brevetoxin, and any remaining binding sites in the wells we re blocked. Samples and con trols we re then added to the wells and allowed to compete with the plate bound toxin for anti brevetoxin antibodies (from a goat) After one hou r, the wells we re rinsed out, removing the samples and all antibodies except those attached to the plate bound toxin. The ant ibodies remaining in the plate we re then visualized using a horseradish peroxidase (HRP) conjugated secondary antibody (rabbit ant i goat antibodies) and the HRP substrate TMB ( 3,3 ',5,5' Tetramethylbenzidine) Absorbance of the wells was read at 450 nm. Selected avian samples were also evaluated with LC MS in order to confirm the presence of parent toxins (PbTx 1:867, PbTx 2:895, PbTx 3:897) and/or specific brevetoxin metabolite s (PbTx 3 Cys:1018, PbTx 2 Ox Cys:1034). A 0.5g equivalent of the me thanol extracts were diluted with de ionized water and passed through a Strata X SPE (solid phase extraction) column (60 mg, 3 mL, Phenomenex) that had been preconditioned with 100% MeOH, and then 20% MeOH (9 ml each). Columns were rinsed with 4.5 m L 20% MeOH and the samples were eluted with 4 m L 100% MeOH. Samples were then evaporated to d ryness and redissolved in 0.5 mL MeOH. All samples were stored at 20C until further analysis. A ll MS analysis was conducted at the Center for Marine Science, Universit y of North Carolina, Wilmington by using a Waters Alliance 2695 LC (Waters, Milford, MA, USA) coupled to a Micromass ZQ mass spectrometer equipped with an electrospray ionization (ESI) probe (Micromass Inc., UK). All analyses were conducted using electrospray ionization with the probe at 3 kV and 250 o C. Samples were first separated on a YMC JSphere ODS L80, S 4 2.0 mm column (YMC Inc., Wilmington, NC, USA). The solvent gradient wa s composed of
20 acidified (0.3 % acetic acid) ACN/H 2 O with initial 50:50 ACN/H 2 O to 95:5 ACN/H 2 O over 40 min. Parent brevetoxins (PbTx 1:867, PbTx 2:895, PbTx 3:897, PbTx 6:911, PbTx 7:869, PbTx 9:899, PbTx 10:871) and brevetoxin metabolites (Cyst PbTx 2:101 8, Ox Cyst PbTx 2:1034) were monitored at indicated masses. The instrument was calibrated with a standard brevetoxin mix containing PbTx 2 and PbTx 3, obtained from the Center for Marine Science, UNC Wilmington, N.C. Determination of Bloom Conditions The timing and location of recovery for all avian samples were analyzed with respect to local bloom conditions. Numerical abundance of K. brevis cells from water samples collected in the region is available from the F WCC FWRI HAB Monitoring Program Week ly summary reports are provided for bloom conditions for the southwest region of Florida as part of ongoing monitoring work, including near and offshore water sampling locations and results. K brevis concentrations in water samples are determined by hand counting cells using a dissecting microscope. Measurements are reported as a range, and a midpoint of the range is used to estimate cell concentration. Red tide status reports are issued weekly. Cell concentrations are correlated to terms used to descr ibe bloom conditions, or red tide intensity, according to Table 3. Table 3. Terms used for describing K brevis bloom conditions in weekly Red Tide Status reports issued by the FWRI HAB Monitoring Program K. brevis cells per liter seawater Bloom Condi tions No cells detected Not Present 1,000 cells or less Present >1,000 to 10,000 cells Very Low >10,000 to <100,000 Low 100,000 to <1,000,000 cells Medium >1,000,000 cells High
21 RESULTS Overview of Avian Samples One hundred and one birds, rep resenting 23 species, were collected from Tampa Bay to Charlotte Harbor during the K brevis bloom in 2005 (Table 4). Most carcasses were recovered during the summer months (Figure 3), but this was not necessarily a reflection of a red tide related trend as nesting seasons and other factors may have influenced this pattern. Figure 3. The number of birds collected each month for necropsy and brevetoxin analysis during the 2005 K brevis bloom. 0 5 10 15 20 25 30 January February March April May June July August September October November December Number of Birds
22 Species No. Tested No. Positive Black-bellied Plover (Pluvialis squatarola) 1 1 Black-crowned Night Heron ( Nycticorax nycticorax ) 1 1 Black Skimmer (Rhynchops niger) 1 1 Brown Pelican (Pelecanus occidentalis) 10 10 Common Loon (Gavia immer) 1 1 Double-crested Cormorant ( Phalacrocorax auritus ) 18 18 Dunlin ( Calidris alpina ) 1 1 Great Blue Heron ( Ardea herodias ) 4 3 Green Heron (Butorides virescens) 2 2 Herring Gull ( Larus argentatus ) 2 2 Laughing Gull (Larus atricilla) 11 11 Least Tern ( Sterna antillarum ) 15 9 Northern Gannet ( Morus bassanus ) 4 4 Osprey ( Pandion haliaetus ) 4 2 Royal Tern ( (Thalasseus maximus ) 3 3 Ruddy Turnstone ( Arenaria interpres ) 2 2 Sanderling ( Calidris alba ) 11 11 Sandwich Tern ( Thalasseus sandvicensis ) 5 5 Snowy Egret ( Egretta thula ) 1 1 Sora ( Porzana carolina ) 1 1 Yellow Crowned Night Heron ( Nyctanassa violacea ) 1 1 White Pelican ( Pelecanus erythrorhynchos ) 1 1 Willet (Tringa semipalmata) 1 1 TOTAL 101 92 Table 4. Summary data of bird samples collected during the 2005 Karenia brevis bloom. Positive results indicate an individual had at least one tissue result above 5 ng PbTx per gram tissue. The ELISA assay used is not able to detect brevetoxins below this level. A result below the limit of detection (
23 Brevetoxin Analysis of Avian Tissues The digestive tissues tested for each bird included liver, stomach contents or whole stomach, intestinal contents or whole intestine, and gall bladder. Additional tissues were analyzed f or the majority of birds examined, with total body burden results to be investigated separately (Karen Atwood, unpublished). Seve ral birds did not have intestinal samples as a result of decomposition, and gall bladders were not recovere d from small birds or decomposed carcasses. If multiple stomach content samples were taken from a bird, the specific sample referred to is indicated or results are reported as an average of all stomach content samples. A total of 321 tissues were sampled, with brevetoxins detected in 264 (82%) samples. Brevetoxins were detected in at least one gastrointestinal sample for 92 of the 101 birds tested (91%). Sixty one of the 321 avian tissues tested (19%) were found to have a brevetoxin level over 200 ng PbTx/g tissue (Table 5). Thirty four birds (34%) had at least one digestive tissue or gastrointestinal content sample greater than 200 ng PbTx /g tissue Cormorants were the most frequently occurring species in this group with twelve individuals.
24 # tissues % tissues positive % tissues >200 ng/g LOW HIGH MEAN LOW HIGH MEAN LOW HIGH MEAN Cormorants (n=18) 71 92% 27%
25 Gall bladder samples were not available for most of the birds sampled due to the relatively small size of this organ and difficulties removing it from previously frozen or partially decomposed carcasses. Of the 29 gall bladders sampled, none fell below the limit of detection for the ELISA, a finding not observed with the liver, stomach or intestinal samples. Also, a greater proportion of all gall bladder samples had very elevated brevetoxin levels in comparison to the frequency distribution for all othe r tissues sampled (Figure 4). Figure 4 Comparison of the frequency interval for brevetoxin levels in gall bladder samples of marine birds as compared to liver, stomach and intestinal samples.
26 5 ). For the latter group, mean brevetoxin concen tration was greatest for intestinal contents, followed by the mean brevetoxin concentration in gall bladder, liver and stomach contents, respectively. Figure 5 Distribution of brevetoxins in digestive samples from cormorants (n=14), brown pelicans (n= 6) and other avian species (n=9) for which gall bladders were analyzed. Mean brevetoxin levels with standard deviation are shown. 0 200 400 600 800 1000 1200 1400 1600 1800 Cormorants Brown Pelicans All other Cormorants Brown Pelicans All other Cormorants Brown Pelicans All other Cormorants Brown Pelicans All other LIVER STOMACH INTESTINE GALL BLADDER In addition to the seabirds described above, the carcasses of fifteen l east t ern ( S. antillarum ) nestlings and fledglings were recovered from rooftop nesting sites in Pinellas County between mid May and early July in 2005. These chicks ranged in weight from 2g to 27g, and all expired as a result of falling from their rooftop nesting area. Samples for these small chicks fre quently required that all gastrointestinal tissues or all viscera be combined for analysis. Twenty nine l east t ern tissue samples were analyzed, with brevetoxins detected in 13 tissues (45%). All positive tissues were less than 40 ng
27 PbTx/g tissue, with the exception of one gastrointestinal tract sample from a fledgling collected June 19 which was found to have 218.81 ng PbTx/g tissue (Table 6 ). TISSUE # samples % positive Low High Mean Whole Viscera 2 100% 14.41 17.82 16.12 Liver 13 46%
28 analysis of the liver and stomach contents, including toadfish bones, revealed relatively low levels of brevetoxin present (14.56 ng/g and 16.40 ng/g, respectively) with higher levels detected in the i ntestinal content s ample ( 103.13 ng /g). The gall bladder ruptured and was therefore not available for analysis. A s andwich t ern ( T sandvicencis ) was brought to a rehabilitation center from Stump Pass State Park on Manasota Key, FL on June 13, 2005. Thi s bird was emaciated and unwilling to fly, dying within 24 hours of care. A fishing hook with a small amount of fishing line was found in the stomach of the bird during necropsy. T he hook was protruding through the stomach wall causing extensive internal injury. A portion of a Spanish s ardine ( H jaguana ) was also recovered from the terns stomach. The fish was found to be below the limit of detection for the brevetoxin assay, as were the intestines of the tern, and only a small amount of brevetoxin was identified in the liver (12.39 ng/g). Cell counts for K. brevis in water samples collected June 14 from nearby Charlotte Harbor and Gasparilla Sound found no cells present, with higher counts collected further north in Sarasota Bay. A juvenile, male brow n p elican ( P. occidentalis ) with wing injuries was collected June 25 from coastal waters in St. Petersburg. A whole fish (unidentified species) was recovered from the pelicans esophagus during necropsy, with subsequent analysis confirming that the viscer a of this fish contained elevated levels of brevetoxin (Table 7 ). The pelican was also found to have low levels of brev etoxin in its liver (30.83 ng/g) and intestine (31.65 ng/g ) but no neurological symptoms were noted by the rehabilitation handlers.
29 Table 7 Detailed summary of birds found with whole or partial fish in stomach during necropsy. K. brevis counts refer to bloom conditions in area at time of recovery of bird. PbT x level is for whole fish unless indicated otherwise. Species Recovery D ate Location Found Stomach Contents PbTx Level K. brevis levels Ring billed Gull ( Larus delawarensis ) February 20, 2005 Siesta Key, FL Mojarra spp. ( Eucinostomus spp.) 92 4 .4 1 (381.95) M edium to high Double c rested Cormorant ( Phalacrocorax auritus ) May 1, 2005 Casey Key, FL Toadfish ( Opsanus sp.) 16.40 N ot present Least Tern ( Sternula antillarum ) May 20, 2005 Madiera Beach, FL Silverside ( Menidia sp.) 16.22 Not present to low Least Tern ( Sternula antillarum ) May 26, 2005 Madiera Beach, FL Silverside ( M enidia sp ) 13.17 Not present to low Least Tern ( Sternula antillarum ) May 30, 2005 Clearwater, FL Silverside ( Menidia sp.)
30 A juvenile, female d ouble c rested c ormorant was discovered alive on a Pinellas County beach of the G ulf coast on August 19, 2005. Water samples taken the day before from nearby Bo ca Ciega Bay, just north of the mouth of Tampa Bay, found medium levels of K. brevis cells for this area. The cormorant was observed displaying symptoms consistent with toxicosis, including severe ataxia and convulsions. The bird died en route to a rehab ilitation facility and did not receive any treatment. Subsequent necropsy found two partially digested pinfish ( Lagodon rhomboides ) in the cormorants stomach (Figure 6 ). ELISA analysis on the pinfish and digestive tissues of this bird returned high leve ls of brevetoxin contamination (Table 8 ). Figure 6 Juvenile female double crested cormorant ( top left) recovered from coastal area of Tampa Bay after exhibiting signs and symptoms of brevetoxicosis. Partially digested pinfish found in the birds stoma ch ( top right ) tested positive for high levels of brevetoxin. Water samples taken from the area the previous day found medium to high cell counts of K. brevis ( bottom ).
31 On the morning of Augus t 25, 2005, twenty l aughing gulls ( L. atricilla ) and nine royal t erns ( T. maxima ) were discovered dead on Siesta Key in Sarasota County (Figure 7 ). Necropsies were completed on two r oyal t erns and one l aughing g ull and each bird was found to have whole o r partial t hread h erring ( O oglinum ) in their stomach. ELISA analysis on these stomach contents indicated high levels of brevetoxin, and elevated brevetoxin levels were also found in the liver of each bird (Table 8 ). Water samples collected in the regio n that week found low to high levels of K. brevis in Sarasota Bay and surrounding coastal areas. Figure 7 A flock of l aughing g ulls and r oyal t erns discovered dead on the beach of Siesta Key in Sarasota County, FL on August 25, 2005 (left). While the c ause of death is unknown, subsequent necropsy and analysis showed the birds recently fed on thread herring containing high levels of brevetoxins (right).
32 On August 29, 2005, visitors to Tierra Verde encountered two l aughing g ulls in severe distress Both birds displayed severe neurological symptoms including difficulty in righting themselves An observer killed both birds with the objective of ending their suffering. The gulls were found to have thread herring in their stomachs upon necropsy and the fish, which were analyzed whole, were found to have high levels of brevetoxin. The intestinal contents and livers of both birds also were found to have elevated levels of brevetoxin (Table 8 ). Tierra Verde, which is located at the northern shore of the mough of Tampa Bay, was experiencing medium to high counts of K. brevis in the region for the week these gulls were recovered (Figure 8 ). Figure 8 Two laughing gulls (top left) recovered from coastal area of Tampa Bay after exhibiting signs and sy mptoms of brevetoxicosis. Partially digested thread herring ( O. oglinum ) found in the stomachs of both birds (top right) tested positive for high levels of brevetoxins. Water samples taken from the area that week found medium to high cell counts of K. br evis (bottom).
33 Species Date collected Tissue PbTx (ng/g) LC-MS 19 August Liver 197.67 Stomach contents (whole sample from partially digested pinfish) 9988.62 positive Stomach contents (muscle & skin of partially digested pinfish) 4095.08 positive Stomach content (all other stomach contents) 2310.32 positive Intestinal contents 2645.35 positive Gall bladder 718.43 positive Royal Tern 25 August Liver 96.86 Stomach contents (thread herring viscera) 4399.50 Stomach contents (thread herring muscle and skin) 153.69 Royal Tern 25 August Liver 140.87 Stomach contents (thread herring whole) 650.52 positive Stomach contents (other contents present) 733.61 ND Intestinal contents 465.10 ND Laughing Gull 25 August Liver 147.11 Stomach contents (partially digested thread herring and other contents present) 387.43 Intestinal contents 272.89 Laughing Gull 29 August Liver 1355.47 ND Stomach contents (partially digested thread herring) 2215.80 positive Stomach contents (other contents present) 345.17 Intestinal contents 2800.73 positive Gall bladder 897.11 positive Laughing Gull 29 August Liver 1044.38 positive Stomach contents (partially digested thread herring) 915.78 positive Stomach contents (other contents present) 653.74 positive Intestinal contents 2021.28 positive Gall bladder 2099.21 positive Table 8. Distribution and amount of brevetoxin (PbTx) found in the digestive organs and gastrointestinal contents of six marine birds collected from the central west Florida coast during the 2005 red tide event. Each of these birds had whole, identifiable fish in their stomachs at time of necropsy. These fish and other gastrointestinal tissues were found to contain elevated levels of brevetoxin. Double-crested Cormorant Samples indicated were evaluated with LC-MS at UNCW Center for Marine Science in order to confirm the presence of brevetoxins (parent toxins and/or metabolites). Positive indicates confirmation of the presence of brevetoxins without quantification. ND indicates that brevetoxins could not be confirmed by LC-MS, but does not indicate the sample is negative for brevetoxins..
34 Seven birds, in addition to those listed in Table 8, had brevetoxin levels in their stomach and/or intestinal samples greater than 300 ng/g. No fish or evidence of a recent fish meal was noted during necropsy of these birds as all had empty or mostly empty stomachs. While six of these birds are not strictly piscivorous, fish are known to comprise at least some part of the diet for all these birds and so the elevated levels are shown in Table 9 with any actua l stomach contents of each bird noted. Species Date Collected Tissue PbTx (ng/g) LC-MS 20 February Liver 269.70 Stomach Contents (empty, some sand) 574.67 Intestinal Contents 371.86 Sanderling 6 March Liver 1313.33 positive Stomach Contents (empty) 1071.46 positive Intestine (whole) 1363.08 positive 9 March Liver 121.52 Stomach Contents (nematodes only) 442.91 Intestinal Contents 67.14 Gall Bladder 418.13 Green Heron 25 June Liver 227.16 Stomach Contents (empty) 727.22 positive Intestine (whole) 270.76 Green Heron 5 July Liver 295.71 Stomach Contents (empty) 47.72 Intestine (whole) 510.77 positive 10 August Liver 100.73 Stomach Contents (empty) 315.67 Intestinal Contents 40.27 Gall Bladder 314.93 23 September Liver 53.73 Stomach Contents (empty, some shell) 1267.42 positive Intestine (whole) 122.88 Samples indicated were evaluated with LC-MS at UNCW Center for Marine Science in order to confirm the presence of brevetoxins (parent toxins and/or metabolites). Positive indicates confirmation of the presence of brevetoxins without quantification. Black-bellied Plover Table 9. Distribution and amount of brevetoxin (PbTx) found in the digestive organs and gastrointestinal contents of seven marine birds collected from the central west Florida coast during the 2005 red tide event. Each of these birds were found to have very high levels of brevetoxin in their stomach or intestines. Sanderling Double-crested Cormorant Great Blue Heron
35 Brevetoxin Analysis of Fish Tissues Twenty five thread herring ( O oglinum ) were collected from rehabilitation centers and a commercial fishery in the study region between April and Octobe r 2005. Whole fish ranged in length from 13.5 cm to 21.5 cm (total length) and in weight from 20g to 90g. Thread herring tissues sampled included eyes, gills, whole viscera and whole muscle (including skin and scales). All thread herring samples analyze d contained detectable levels of brevetoxins as measured by ELISA. Brevetoxin levels in O. oglinum tissues ranged from 24.27 ng PbTx/g tissue to 5839.89 ng PbTx/g tissue. Highest levels were found in the viscera as compared to eyes, gills or muscle of a fish (Table 1 0 and Figure 9 ). Sixteen of the 25 viscera samples (64%) had over 1000 ng PbTx per gram tissue, and only one viscera sample was less than 200 ng PbTx per gram tissue. Muscle tissue samples contained the lowest levels of brevetoxins as compa red to other tissues sampled, with only one muscle tissue greater than 200 ng PbTx per gram tissue. Monthly mean concentrations are shown in Figure 9 Two of the twenty five thread herring sampled were gravid females. For these two individuals, eggs we re analyzed separately from other viscera and contained 585.42 ng PbTx per gram eggs and 843.17 ng PbTx per gram eggs. In both of these cases the brevetoxin levels for eggs were less than the toxin levels in the whole viscera sample of the same fish (3617 .02 ng PbTx/g tissue and 3269.90 ng PbTx/g, respectively). Only three of the fifteen tissue samples taken from four individual finger mullet ( Mugil spp.) were found to have detectabl e levels of brevetoxin (Figure 10 ). Mullet ranged in length from 15cm to 17cm (total length), and in weight from 36g to 42g. For those tissues with detectable levels, all were below 15 n g PbTx per gram tissue (Table 10 and Figure 10 ). Al l 39 tissues sampled from nine S panish sardines ( S auritus ) and one scaled sardine ( H jaguana ) contained detectable levels of brevetoxins (Table 10 and Figure 10 ). Whole sardines ranged in length 11.6cm to 18.5cm (total length) and in weight from 17g to 57g.
36 Figure 9 Brevetoxin levels detected in locally caught thread herring ( O oglinum ) obtained from a local commercial fishery and the food supply of participating rehabilitation facilities. All brevetoxin results were determined by ELISA. Tick marks indicate the LD 50 for oral ingestion of brevetoxins in mice (520 ng PbTx/g mous e) and the level at which oyster beds are closed for human consumption (800 ng PbTx/g oyster tissue). 0 500 1000 1500 2000 2500 3000 Eyes Gills Viscera Muscle+skin Tissues PbTx (ng/g) April (n=4) May (n=3) June (n=6) July (n=2) August (n=3) September (n=5) October (n=2)
37 Figure 10 Brevetoxin levels detected in mullet ( Mugil spp.), S panish sardines ( S auritus ) and a scaled sardine ( H jaguana ) obtained from a local c ommercial fishery and the food supply of participating rehabilitation facilities. All brevetoxin results were determined by ELISA. Tick marks indicate the LD 50 for oral ingestion of brevetoxins in mice (520 ng PbTx/g mouse) and the level at which oyster beds are closed for human consumption (800 ng PbTx/g oyster tissue). 0 200 400 600 800 1000 1200 1400 1600 Eyes Gills Viscera Muscle+Skin Tissues PbTx (ng/g) Finger Mullet (June, n=2) Finger Mullet (August, n=2) Spanish Sardine (May, n=2) Spanish Sardine (June, n=2) Spanish Sardine (July, n=3) Scaled Sardine (August, n=1) Spanish Sardine (September, n=2) low high mean low high mean low high mean Eyes 60.47 259.04 125.53 15.51 155.89 60.16
38 harengulus ), and thread h erring fry Minnow samples represent whole, m acerated fish as opposed to individual tissue samples. In some cases two or three individuals were pooled and tested as a single sample (Figure 11 ). All samples tested were below 500 ng PbTx/g fish. Figure 1 1 Brevetoxin levels in small minnows caugh t along the west central coast of Florida during the 2005 red tide. All brevetoxin results were determined by ELISA. 0 50 100 150 200 250 300 350 400 450 500 Whole Fish PbTx (ng/g) Bay Anchovy (May, n=2) Sheepshead Minnows (May, n=2) Anchovy spp. (June, n=4) Thread Herring fry (June, n=1) Silver Jenny (June, n=1) Bay Anchovy (July, n=5) Bay Anchovy (September, n=3) Twelve dead fish observed being actively scavenged along beaches during the 2005 bloom were collected, dissected and analyzed for br evetoxin content. Birds observed feeding upon red tide killed fish were generally limited to shorebirds and gulls. Video recordings of scavenging behavior were taken, and a tendency for shorebirds such as ruddy turnstones, sanderlings, and gulls to bore into the eyes, gills and intracoelomic cavity while scavenging was noted. Subsequent dissection of beached fish verified the preferential feeding on eyes, gills and viscera by scavenging birds, with many fish lacking eyes, gills and/or internal organs (Fi gure 1 2 ). Only eight of the twelve beached fish were recovered with eyes intact, and all of the eyes sampled contained detectable levels of brevetoxin. Results for eyes ranged from 55.46 ng/g to 929.43 ng/g, with six of the eight samples (75%) over 400 n g PbTx per gram eye tissue (Figure 1 3 and Table 11 ).
39 Figure 1 2 Sanderlings observed actively scavenging red tide killed fish along a beach on the west central coast of Florida during the 2005 Karenia brevis bloom (left). Evidence of shorebird prefer ence for eyes, gills, and internal organs during scavenging of beached fish (right). Three of the twelve beached fish had no recoverable gill tissue remaining. Of the ten gill tissue samples, all had detectable levels of brevetoxins and four (40%) were greater than 1000 ng/g. All viscera samples were also found to have very high levels of brevetoxin, with results ranging from 150.15 ng/g to 95,753.08 ng/g. It should be noted the low finding of 15 0.15 ng/g was from a mullet ( Mugil spp.) whose liver had been scavenged through a borehole and was therefore absent in the sample. One thread herring was also found upon dissection to have no viscera present as it had apparently been removed entirely via an opening near the gill region. Brevetoxins were detected in all samples of muscle tissue, with results generally lower than levels found in the viscera and gill tissues of the same fish (Figure 1 3 and Table 11 ).
40 Figure 13 Brevetoxin levels in beac hed fish collected from beaches along the west central coast of Florida during the 2005 K brevis bloom. All brevetoxin results were determined by ELISA. Tick marks indicate the LD 50 for oral ingestion of brevetoxins in mice (520 ng PbTx/g mouse) and the level at which oyster beds are closed for human consumption (800 ng PbTx/g oyster tissue). 95,753.08 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 11000 12000 Eyes Gills Viscera Muscle & Skin Tissues PbTx (ng/g) 0 20000 40000 60000 80000 100000 120000 Mullet (February) Thread Herring (February) Scaled Sardine2 (March) Mullet (May) Mullet 1 (June) Scaled Sardine1 (June) Scaled Sardine2 (June) Thread Herring1 (June) Thread Herring2 (June) Mullet 2 (June) Thread Herring 3 (June) Scaled Sardine1 (March)
40 Low High Mean Low High Mean Low High Mean Eyes (n=8) 431.61 929.43 644.05 429.73 856.38 643.06 55.46 494.23 236.97 Gills (n=9) 855.98 1251.65 1067.27 735.56 1119.04 927.30 116.40 1007.08 407.94 Viscera (n=11) 2127.90 8726.46 5155.55 2586.62 95753.08 27488.64 150.15 685.08 479.84 Muscle (n=12) 302.03 654.77 437.62 158.40 692.17 392.54 31.63 125.7 62.77 Thread Herring (n=4) Scaled Sardines (n=4) Mullet (n=4) Table 11. Brevetoxin levels in beached fish collected from beaches along the west central coast of Florida during the 2005 Karenia brevis bloom. All brevetoxin results were determined by ELISA and reported as ng PbTx per gram tissue.
41 DISCUSSION Brevetoxins in Digestive Tissues of Marine Birds Results from this study indicate that exposure to brevet oxins is widespread among piscivorous marine birds during K brevis blooms O ne or more digestive tissue s and/or gastrointestinal content sample s contained a detectable level of brevetoxin for more than 90% of the birds sampled Brevetoxins were detected in at least one tissue for all 23 species of marine birds that were represented in this study, showing that exposure can occur among a variety of birds occupying various feeding niches in the region. Results of brevetoxin analyses for the stomach, intes tine, liver and /or gall bladder of sampled birds support the theory that ingestion is the primary route of exposure for these animals during red tide blooms A t oxicokinetic study of rats exposed to PbTx 3 via ingestion found that the liver concentrated t he highest levels of brevetoxins (Cattet and Geraci, 1993), whereas Benson, et al. (2005) determined that no accumulation of brevetoxin or metabolites occurs in the liver of rats repeatedly exposed to toxins via inhalation. In the results presented here 83 % of the liver samples from all birds sampled were positive for brevetoxins. One hundred percent of the avian gall bladders sampled in this study were positive for brevetoxins and the gall bladder frequently had the highest mean brevetoxin levels of tissues sampled in brown pelicans and cormorants for which gall bladder s were available. Furthermore, in a comparison of all liver, stomach, intestinal and gall bladder samples, a greater proportion of the gall bladder samples was found to have higher (>5 00 ng/g) levels of brevetoxin than the proportion of other tissues. The mean gall bladder levels were greater than the mean levels in the liver, stomach or intestinal samples for representative cormorants and pelicans, but this trend was not consistently seen in all species or individuals for which gall bladders were sampled. This is likely indicative of metabolic processes as ingested brevetoxins move through the gastrointestinal tract, and also due to differences in the time elapsed between exposure and death in individuals
42 sampled It may also, however, provide indications to the exposure sensitivity of different avian species to brevetoxins. A more complete analysis of brevetoxin levels in avian tissues will be presented in greater depth at a later d ate (Karen Atwood, unpublished). Poli, et al. (1990) and Kennedy, et al. (1992) concluded that the hepatobiliary system was largely responsible for metabolism and excretion of brevetoxins in toadfish ( O. beta ) and rats following intravenous administration. Th e results presented here offer evidence that the biliary system of birds also may be an important pathway for metabolism and excretion of ingested brevetoxins. Biotransformation of toxins occurs in the liver, which produces bile that is received by th e gall bladder (Berne and Levy, 2000). Feeding stimulates the release of bile via the bile duct into the small intestine and, during periods of fasting, bile is stored and concentrated in the gall bladder (Bowen, 2001). Toxins that pass to the intestine from the gall bladder may be eliminated in the feces or reabsorbed (Duffus and Worth, 2001). Woofter, et al. (2005) suggested that brevetoxin is reabsorbed by the intestines after biliary secretion, and that this would account for the sustained blood leve ls observed in mullet many days after aqueous exposure. The elevated gall bladder levels observed in birds sampled during this study may reflect a situation in which these birds ingested a sublethal amount of brevetoxin, but were behaviorally impaired and unable to feed as they metabolized the toxin, resulting in storage and concentration in the gall bladder. The birds sampled here obviously did not survive this scenario if in fact it was their experience. It is possible that some birds are exposed to hig h, but sublethal, doses of brevetoxins during a meal, are then unable to feed during metabolism and recovery, but eventually recover after a period of fasting. This raises questions regarding the toxicity of the brevetoxins and derivatives in the post exp osure bile that is excreted once feeding is resumed, and the potential effects if there is reabsorption in the small intestine. While all species examined were found to have some detectable level of brevetoxins in at least one digestive tissue or gut conte nt sample, differences could be seen among families of birds. Double crested c ormorants (Family: Phalacrocoradae) are considered one of the avian species most impacted when red tides occur in the region (Quick and Henderson, 197 4 ; Forrester, et al. 1977 ; Kreuder, et al., 2002 ) In this study,
43 cormorants were the most abundant species sampled (Table 4) and were found to have a greater percentage of tissues with more than 200 ng PbTx/g than any other sea bird species sampled Cormorants are abundant in F lorida ( Vidal Hernandez and Nesbitt, 2002 ) They are efficient predators and considered opportunistic generalists, diving from the surface to pursue schooling or, occasionally benthic fish and generally feed ing on prey that is most abundan t (Wires, et al 2001; Elphick, et al. 2001). During beach surveys in 2005, cormorants could often be seen foraging along local beaches even during intense onshore bloom conditions when b rown p elicans gulls and other species were absent (Figure 1 4 ) It is possible th at the opportunistic nature o f these birds could draw them to the erratic swim patterns of brevetoxin intoxicated fish. N eurointoxicated f ish are potentially easy prey as they often display behavior s attractive to predators, such as floundering at the sur face and a depressed flee response (Quick and Henderson, 1974). Figure 1 4 Double c rested c ormorants congregated along Casey Key Beach in Sarasota during onshore bloom conditions on March 3, 2005. These birds appeared to be foraging alongshore while gul ls, terns and pelicans were absent.
44 The majority of cormorants examined were juveniles (85%) 67% were found to have gastrointestinal nematode infestations, and 38% were noted to be very thin or emaciated (Appendix I lists bo dy weight information) This is consistent with observations made by Kreuder, et al., (2002), as a high occurrence of endoparasitism and immaturity was observed in cormorants with suspected brevetoxicosis. Gastrointestinal parasites are not uncommon in w ild animal s but it is possible that a heavy parasite load combined with brevetoxin exposure may overwhelm the immune system of young or weak bird s making an other wise survivable condition more serious. Pelicans, ospreys and gannets were not found to con tain brevetoxins in their gastrointestinal tissues at levels as high as cormorants, gulls and terns, herons and egrets, and shorebirds ( Table 5) Perhaps most noteworthy was the difference between mean brevetoxin concentrations in tissues from b rown p elic ans versus d ouble c rested c ormorants. The r e are similarly sized, large populations of both these bait fish consuming species in Florida (Table 2) The r esults offered here i ndicate that despite similarities between the two populations, there could be imp ortant differences in their exposure to or tolerance of brevetoxins during blooms of K. brevis in the region. P ossible reasons why b rown p elicans may be less vulnerable to brevetoxin exposure during red tide blooms as compared to d ouble crested c ormorants include different physiological tolerances to brevetoxins, or behaviors designed to reduce exposure or vulnerability to toxins, such as prey avoidance or regurgitation of contaminated prey. Also, interspecific differences in body burdens of parasites or other contaminants which could lead to secondary effects following brevetoxin exposure could play a role As there are numerous, complex factors that can influence fitness in a population, a more thorough examination was beyond the scope of this study. In addition to cormorants, other avian families from which individuals were found to have elevated levels of brevetoxins were g ulls and terns (Family: Laridae) herons and egrets (Family: Ardeidae ), and shorebirds ( Family: Scolopacidae ). Members from t hese avian familie s consume a variety of foods and forage in a number of diverse ways Gulls are generally considered opportunistic omnivores, while terns and skimmers have more specific methods of prey capture and can often be observed foraging on small fish close to shorelines and along beaches ( Elphick et al., 2001 ) In a comparison of tissue
45 analyses among l arids, the highest levels of brevetoxins were found in laughing g ulls and royal t erns. These two species are often seen together in flocks on be aches and the recovery and stomach content findings from the r oyal t erns and l aughing g ulls on Siesta Key in August 2005 suggest s that they forage together or more likely, that laughing gulls are pirating terns successfully as they forage (Elizabeth F orys, personal communication) Herons and egrets are found in both fresh and marine habitats, and eat primarily fish, but will also feed on crustaceans, amphibians and occasionally small rodents or the chicks and eggs of other birds ( Elphick, et al. 20 01). They are wading birds and forage in shallow coastal waters canals or along mangrove fringe in search of small fish and other prey. The tissue analyses of birds in this family, and the high levels of brevetoxins found in the g reen a nd great b lue h er ons (Table 9 ) indicate ardeids are not avoiding marine prey during red tide blooms in favor of a freshwat er diet. However, it cannot be concluded that fish were the source of brevetoxin exposure for these birds as fish were not confirmed in stomach conte nts during necropsy, and due to the variety in their diet, exposure may have resulted from consumption of shellfish or other prey items S anderlings, turnstones, d unlins and w illets (Family: Scolopacidae ) along with a variety of plovers (Family: Charad ridae), generally forage for invertebrate prey such as interstitial insects, amphipods, and polychaete worms, in the intertidal zone of beaches ( Elphick, et al. 2001) Some of their prey items, including small bivalves (eg., Donax spp. ) are known to ac cumulate brevetoxins (Cummins, et al. 1971; Roberts, et al. 1979) Mass die offs of invertebrates in the intertidal zone h ave also been documented immediately following red tide s in the region ( Gunter, et al. 1947; Simon and Dauer, 1972 ) and the absen ce of a number of species which contribute to the diet of shorebirds, such as polychaetes and amphipods, was personally observed on beaches during the 2005 bloom. Therefore, confirmation of shorebirds scave nging red tide killed fish is important because i t supports the possibility that shorebirds switch their diets from traditional sources due to contamination or reduced availability to abundantl y available dead fish during red tide blooms Fish deposited onshore could vector brevetoxins to shorebird s as well as gulls and other scavenging animals, if toxins have bioaccumulated in the fish prior to their
46 death and deposition This is an aspect of the food web not generally considered as fish typically do not comprise a large portion of the diet of smal l shore birds Based on my observations, beached fish do appear to be an important alternative food source during red tides. Samples from dead fish being scavenged by shorebirds found some alarmingly high levels of brevetoxin s in the parts of fish known t o be targeted by shorebirds such as the eyes, gills and viscera As further evidence that red tide killed fish may contain high levels of toxins, there were also anecdotal accounts of terrestrial organisms, including bobcats, rats and pet dogs, exhibitin g signs of neurointoxication following scavenging of fish on red tide impacted beaches in 200 5 and past red tides ( Wildlife Center of Venice, personal communication ; Ernst, 2003 ) Most of the beached fish sampled particularly the s caled s ardine with 95, 000 ng/g brevetoxin in the viscera, would appear to contain lethal levels of toxin for a small shorebird weighing less than 60g. When Aldrich and Ray (1965) exposed two similarly sized chicks to a total ingestion of 198,000 ng PbTx, an amount comparable t o what could be consumed from the Scaled s ardine by a s anderling assuming a consumption of 2g viscera both chicks lost equilibrium within 10 hours and expired within 22 hours. Elevated levels of brevetoxins were found in the digestive tissues and conten ts of several shorebirds, but symptoms associated with brevetoxicosis such as ataxia and spastic movements (Kreuder, et al. 2002) were not noted in birds provided by rehabilitators. Also, there were no reports of large numbers of dead or dying shorebirds in the hours or days following observed consumption of red tide killed fish on beaches in February and March of 2005 A morbidity and mortality event of primarily scolopocids ( s anderlings, r uddy t urnstones and w illets) was reported and observed in the f all of 2005 from Tampa Bay to Charlotte Harbor, nine months after the start of the red tide bloom ( Suncoast Seabird Sanctuary and Wildlife Center of Venice, personal communication ) Individuals collected by rehab ilitation facilities were reported as ex hib it ing signs and symptoms of severe gastrointestinal distress, lower body paralysis and systemic disease. T hese birds do not consume fish whole and therefor e the nature of their diet was no t able to be confirmed in stomach contents during necropsy Brevet oxins were detected in the digestive tissues of all of these birds, but very h igh levels of brevetoxins were not
47 consistently found in all shorebirds associated with the mortality event (Appendix II) In the shorebird s found to have elevated brevetoxins i n their tissues i t could not be confirmed that th is was the result of scavenging contaminated fish rather than exposure through traditional invertebrate prey. During the course of this study, though, we did document that shorebirds are consuming dead fis h onshore during red tide blooms, and that those fish frequently contain elevated levels of brevetoxin s in the tissues targeted by scavenging birds A few species of birds for which mortalities have been reported during past K. brevis blooms were not wel l represented during in this study. Red breasted m ergansers ( Mergus merganser ) and l esser s caup ( Aythya affinis ) were completely absent among the collected bird carcasses despite reports that these species were heavily impacted during past blooms ( Quick a nd Henderson, 1975; Forrester, et al. 1977). O nly one white p elican ( P. erythrohynchos ) was recovered during this study as compared to the 2001 bloom in the region during which time dozens of these birds were treated by local rehabilitators for neurologi cal symptoms ( Pelican Man s Bird Sanctuary, personal communication ). As previously described, the 2005 bloom reached onshore waters in February and persisted through the spring and summer months. In contrast, t he 1974 red tide occurred from approximately October to May and the 2001 red tide lasted from about September to early January. These observation s underscore the potential importance of bloom timing as an influence on which avian species are affected when a bloom occurs A bloom which initiates in the late fall and persists through the winter is likely to c ontribut e to increased mortality and morbidity among Florida s migratory bird species that occur in winter and early spring It should be noted that least terns, which were included in this stud y, are a migratory bird occurring in Florida during their nesting season from April to September. Prey Fish and Brevetoxins in Stomach Contents of Birds An important objective of this study was to determine the role of prey fish in vectoring brevetoxins to marine birds. Brevetoxins were consistently found in the digestive tissues or gastrointestinal contents of all species and most individual birds examined, and samples of the small schooling fish targeted by piscivorous marine birds
48 were also shown to carry potentially harmful levels of brevetoxins. In order to make the link between brevetoxin containing prey fish and their avian predators, it was important to find birds whose last fish meal, and the brevetoxin levels therein, could be confirmed throug h stomach content analysis. Six of the 101 seabirds examined were found to have whole or partial fish in their stomachs which contained elevated levels of brevetoxins. Of these, three birds had exhibited signs and symptoms consistent with brevetoxicosis b efore dying, an indication that their fate directly resulted from poisoning. While toxin levels for sublethal versus lethal exposures for brevetoxins in marine birds are not known, it is possible that the levels found in the d ouble c rested c ormorant colle cted on August 19, 2005 and the l aughing g ulls recovered on August 29, 2005 (Table 8) are somewhere in the range of an acute exposure. Also, finding brevetoxin containing fish in the stomachs of these birds confirms that piscivorous birds are feeding on c ontaminated fish during K. brevis blooms and not all individuals are able or willing to avoid red tide impacted fish entirely. Tester, et al. (2000) documented the transfer of brevetoxins from copepods to the zooplanktivorous pinfish ( L. rhomboid e s ), an d here we present evidence of consumption of brevetoxin containing pinfish by a d ouble crested c ormorant. These findings offer confirmation of vectorial transport of brevetoxins from dinoflagellate ( K. brevis ) to predatory marine birds via planktivorous f ish an ecologically important link in the Gulf of Mexico food web. Eight of the 15 least t ern chicks sampled were found to have small minnows in their stomachs. Unlike adult seabirds, chicks have their meals brought to them by adults and do not rely on being able to successfully forage for themselves. Analysis on these small minnows and gastrointestinal contents and tissues did not reveal high levels of brevetoxins, with the exception of one gastrointestinal sample for a single fledgling. While nine of the fifteen chicks tested were positive for brevetoxins, result s were generally low (Table 6 ) and do not indicate that brevetoxin exposure was an important contributor to their loss. For the remaining 83% of the marine birds examined for this study, the re was no evidence of a recent fish meal in their stomach contents. This suggest s that acute, lethal poisoning soon after feeding may not be the fate of most piscivorous seabirds during
49 blooms of K. brevis These results could be more consistent with bir ds being exposed to chronic low levels in their diet via contaminated prey, or acute exposures that result in prolonged illness rather than sudden death. Consumption of brevetoxins via prey fish could potentially impair immune and behavioral functioning i n marine birds triggering a cascade of events in which functional behavior of an individual, such as feeding, preening, and injury aversion, is impaired due to ingestion of contaminated prey ( Elphick, et al. 2001; Kreuder, et al. 2002). An investigatio n into a mortality event of dolphins in 1987 1988 supports such a scenario. Here it was determined that many individuals died as a result of opportunistic infection or other factors that were only fatal due to the physiological weakening of individuals, p otentially due to brevetoxin exposure (Anderson and White, 1992). During this study, there were two examples of birds potentially incapacitated such that survival behaviors may have been compromised. The gulls and terns collected from Siesta Key Beach on August 25, 2005 were found to have recently fed on t hread h erring, and for all three birds sampled, those fish contained elevated levels of brevetoxins (Table s 7 and 8 ). While these results offer important information about brevetoxins in the fish that t hese birds are feeding on during red tide conditions, it does not appear the contaminated fish were the direct cause of the mortality event. There is strong evidence from weather reports and necropsy findings including sloughing and discoloration of musc le tissue and degradation of intestinal tissues, that these birds were the victim of a lightening strike. As previously described, b revetoxin exposure in birds is associated with "drunken" behavior and a reduced flight response. It is possible that the c ontaminated fish the gulls and terns recently fed on resulted in the ir not responding normally to the storm and threat of lightening thereby contributing to the ir fate The second example is t he d ouble c rested c ormorant recovered from Sanibel Island on Au gust 2, 2005 This individual may also have been experiencing reduced functional behavior as a result of toxin exposure This adult bird was hit by a car and euthanized by local rehabilitators due to spinal injuries. Analysis on the gastrointestinal tis sues detected 719.25 ng PbTx per gram gallbladder tissue and 28.46 ng/g, 32.08 ng/g, and 31.66 ng/g in the liver, stomach contents, and intestinal contents respectively. This toxin profile of elevated brevetoxins in the gallbladder as well as lower levels in all other
50 gastrointestinal tissues is consistent with the tissue profiles found in other cormorants collected which were labeled as suffering from red tide syndrome (ataxia, slow blink, head tremor) prior to expiring (Appendix III) Therefore, this bi rds nervous system could have been compromised to the point that it was incapable of an appropriate flight response when threatened by an oncoming vehicle. Brevetoxins in Common Prey Fish of Marine Birds Results from fish tissue analyses suggest that pre y fish can bioaccumulate brevetoxins during K. brevis blooms, and piscivorous marine birds may be vulnerable to toxin exposure via their diet. A mong the live caught fish sampled, t hread h erring were found on average to have higher levels of brevetoxins as compared to the sardines and juvenile mullet sampled (Table 1 0 ) The levels of brevetox ins detected in the viscera of t hread h erring were extremely high and would appear to pose a risk to predatory marine birds as they were similar to the levels in fish recovered from the stomachs of the d ouble crested c ormorant and laughing g ulls which exhibit ed symptoms of brevetoxicosis prior to expiring (see Tables 8 and 1 0 Figure 8). If it is assumed that birds are not effectively avoiding or regurgitating breveto xin containing prey a potential exposure level from contaminated fish can be calculated simply from the findings and known consumption rates of certain species (Lefebvre, et al. 2002). For example, a thread h erring taken from the food supply of the Wild life Center of Venice in June was found to have 1892 ng PbTx per gram viscera, and a total viscera weight of 6.9 grams, which calculates to a total body burden of more than 13,000 ng PbTx in this 90 gram fish. A 45g t hread h erring collected in September w as found to have 2231 ng PbTx per gram viscera, and a total viscera weight of 4.7 grams, resulting in a body bur den of at least 10,000 ng PbTx for the whole fish Reports of d aily consumption rates for double crested c ormorants vary from an estimate of 16 % of their total body weight ( Schramm, et al., 1987 as referenced in Bayer, 1989) to the more common estimate of approximately a pound ( 500g ) of fish per day ( USFWS, 2007). In c onsidering only the toxin levels of the viscera and ignoring all other tissue levels, consumption estimates for birds feeding on 250 500g of heavily contaminated thread herring could be estimated at over 3 5 ,000 ng PbTx/day
51 The s ardines sampled, whil e not generally as high as the thread h erring, were also shown to carry a body bur den of toxins with higher concentrations in the viscera (Table 1 0 ) The juvenile mullet and small minnows collected nearshore were found to have relatively low or no detectable toxin and may pose less of a risk to those birds targeting these small species This theory is supported by the l ess frequent occurrence of high brevetoxin levels found in the gastrointestinal tissues of birds known to feed exclusively on small minnow species nearshore and along beaches, such as least t erns (Table 6 ) and black s kim mers. Management Implications The se r esults provide evidence that small schooling fish such as thread h erring and sardine species, which represent an important source of food for marine birds on the central west Florida coast, can contain elevated levels of brevetoxins B revetoxin s were consistently found in the tissues of live caught fish with the highest levels detect ed in the viscera of t hread h erring. Facilities utilizing whole bait fish from Floridas Gulf coast for animal consumption should be aw are of the implications of these findings. Rehabilitation centers in the region, as well as a quariums, zoos and amusement parks with animals consuming baitfish, may need to consider the potential for these fish to carry potentially harmful levels of breve toxins (Anderson and White, 1989; Krueder et al. 2002). Facilities may consider importing baitfish from unaffected a reas during bloom conditions Also, there are implications for the practice by fisherman of throwing fish offal and viscera discards to b egging seabirds during fish cleanup. Evidence presented here indicates this may pose a health concern for these birds. In general, t here are indications that red tide blooms caused by harmful algae may be increasing in frequency and/or impacting larger or novel coastal areas worldwide ( Hallegraeff. 1993 ; Van Dolah 2002; Smayda, 2002). I n late September 2007, a K brevis bloom began affec ting coastal areas of northeast Florida, a region not generally prone to these blooms. Educational efforts regarding r isks to marine birds particularly potential exposure dynamics in their food supply and possible treatment methods would be useful to rehabilitators and others not accustomed to dealing wit h brevetoxin impacted animals. Species recovery plans for threaten ed and endangered sea or shore birds should
52 also consider the potential risk of spati al or temporal changes to K. brevis blooms. Increasing frequency or intensity of blooms may result in greater toxin exposure or impacts on prey species abundance for mari ne birds. Recommendations for Future Investigations While this study shows that brevetoxins can be found in the gastrointestinal contents of marine birds and the fish they prey upon there is a lack of applicable clinical information for dose response lev els of brevetoxin ingestion in birds Therefore, we cannot conclusively say that consumption of brevetoxin contaminated prey directly led to the illness or death of individuals Any exposure to marine toxins may tip the balance for animals subject to com petition for food, loss of habitat, parasites, disease, and increasing boat traffic as was determined during past investigations of dolphin mortalities in Florida (Anderson and White, 1992) Ingestion of chronic, subacute levels of brevetoxins, or sporad ic acute exposures, could impair critical behaviors of birds, including feeding, preening and injury avoidance, and cause individuals to succumb more easily to other factors, such as parasites, disease or injury. Future studies investigating the impact of sublethal ingestion of brevetoxins on immune and behavioral function would improve our understanding how K. brevis blooms impact marine birds. Also, a long term study tracking fluctuations in levels of brevetoxins in avian populations during non bloom and bloom periods would help illustrate potential changes in exposure s or the number and species of birds impacted over time. Long term ecosystem models of baitfish toxicity and abundancy during red tide blooms could also be useful to understanding impacts on avian predators Thread h erring findings presented in this study indicate the ability of this species to accumulate brevetoxins in their viscera and potentially vector those toxins to predatory marine birds and other organisms It is also important to point out that this species was unavailable from commercial fisheries in the later months of 2005 due to purported red tide related losses. Gulf coast landings of Atlantic t hread h erring comprise approximately 88% of the fishery for the state of Florida, and the 2005 total landings were reduced by 22% from historical averages from 1982 (FWRI Species Account, 2006). This dramatic reduction
53 in availability may result in avian stress and disease due to reduced prey availability or a reduction in potentiall y contaminated fish could also drive predatory birds from affected areas and result in fewer incidences of exposure. T he impact on organisms that scaveng e red tide killed fish merits further consideration. There are implications for both terrestrial and pelagic scavengers, including sharks which have been documented gorging on floating dead and decomposing fish (Steidinger and Ingle, 1972). T he shorebirds of Floridas Gulf coast face numerous pressures as coastlines are urbanized and beaches become incr easingly crowded with human traffic. It is important to consider all influences on the health of shorebird populations in order to determine appropriate conservation measures. In vestigation s into the abundance and toxicity of traditional food sources for shorebirds during bloom and non bloom periods and the potential risks or benefits associated with switching to dead fish during red tide blooms would greatly improve our understanding of how K. brevis blooms in this region impact resident and migratory shorebird populations Finally, it is possible that periodic red tide events along the WFS are the kind of environmental perturbation that contributes to the diversity of avian populations and other marine organism s (Va rgo, 1987). Double crested c ormora nts have historically stood out as a species impacted by red tide blooms in the area (Quick and Henderson, 1974; Forrester, et al. 1977; Emslie, et al., 1996; Kreuder, et al., 2002 ). In other regions, these birds are considered nuisance species for their adept fishing skills and consumption rates, prompting harassment measures and calls for population control by local fishermen (Bayer, 1989 and references therein; Wires, et al. 2001). Cormorants also cause some of the most d ramatic impacts on vegetation at nesting colonies, resulting in impacts on other birds due t o competition for nesting space and habitat degradation (Wires, et al. 2001). Perhaps red tide blooms in the southwest region of Florida may serve to limit populations of more competitive bir ds, such as cormorants, creating opportunities for other species to flourish A long term investigation into fluctuations of avian populations in context of red tide blooms may offer insight into this possibility.
54 REFERENCES Anderson, D.M. 1994. Red Tides. Scientific American 271(2): 62 68. Anderson, D.M., and A.W. White ( Eds. ) 1989. Toxic dinoflagellates and marine mammal mortalities: Technical Report. Woods Hole Oceanographic Institute. WHOI 89 36. Anderson, D.M. and A.W. White. 1992 Marine biotoxins at the top of the food chain. Oceanus 35(3): 55 61. Anderson, D.W., Gress, F. and Mais, K.F. 1982. Brown pelicans: influence of food supply on reproduction. Oikos 39: 23 31. Armstrong, I.H., J.C. Coulson, P. Hawkey, and M.J. Hu dson. 1978. Further mass seabird deaths from paralytic shellfish poisoning. British Birds 71: 58 68. Baden, D.G. and T.J. Mende. 1982. Toxicity of two toxins from the Florida red tide marine dinoflagellate, Ptychodiscus brevis Toxicon 20: 457 461 Baden, D.G., A.J. Bourdelais, H. Jacocks, S. Michelliza and J. Naar. 2005. Natural and derivative brevetoxins: historical background, multiplicity, and effects. Environmental Health Perspectives 113(5): 621 625. Bayer, R. D. 1989. The cormorant/f isherman conflict in Tillamook County, Oregon. Studies in Oregon Ornithology No. 6. 95 pp. Benson, J.M, F.F. Hahn, T.H. March, J.D. McDonald, A.P. Gomez, M.J. Sopori, A.J. Bourdelais, J. Naar, J. Zaias, G.D. Bossart and D.G. Baden. 2005. Inhalation to xicity of brevetoxin 3 in rats exposed for twenty two days. Environmental Health Perspectives 113: 626 631. Berne, R.M and M.N. Levy (Eds.). 2000. Principles of Physiology, 3 rd Edition. Mosby, Inc., St. Louis, MO. 680 pp. Bossart, G.D., D.G. Baden, R.Y. Ewing, B. Roberts, and S.D. Wright. 1998. Brevetoxicosis in manatees ( Trichechus manatus latirostris ) from the 1996 epizootic: gross, histologic, and immunohistochemical features. Environmental Toxicologic Pathology 26: 276 282.
55 Bourdelais, A.J ., C.R. Tomas, J. Naar, J. Kubanek and D.G. Baden. 2002. New fish killing alga in coastal Delaware produces neurotoxins. Environmental Health Perspectives 110(5): 465 470. Bowen, R.A. 2001. Pathophysiology of the Digestive System. Colorado State Un iversity Hypertextbook. Available online at: http://arbl.cvmbs.colostate.edu/hbooks/pathphys/digestion/index.html Last accessed October 2007. Bretz, C.K., K. Thomas R. Kvitek, and F. Barry. 2000. Influence of harmful algal blooms on the foraging behavior of shorebirds in central California. Presented at Symposium on Harmful Marine Algae in the United States, Woods Hole Oceanographic Institute. Burger, J. and M. Gochfeld. 1991. Human activity influence and diurnal and nocturnal foraging of sanderlings ( Calidris alba ). The Condor 93(2): 259 265. Butcher, G.S. and D.K. Niven. 2007. Combining data from the Christmas Bird Count and Breeding Bird Survey to d et ermine the c ontinental status and trends of North America b irds National Audubon Society, Technical Report. 34 pp. Available online at: http://www.audubon.org/bird/sta teofthebirds/CBID/content/Report.pdf Last accessed October 2007. Castonguay, M ., M. Levasseur, J.L. Beaulieu, F. Grgoire, S. Michaud, E. Bonneau, S.S. Bates 1997. Accumulation of PSP toxins in Atlantic mackerel: seasonal and ontogenetic variations Journal of Fish Biology 50: 1203 1213. Cattet, M. and J.R. Geraci. 1993. Distribution and elimination of ingested PbTx 3 in rats. Toxicon 31: 1483 1486. Colbert, A.A., G.I. Scott, M.H. Fulton, E.F. Wirth, J.W. Daugomah, P.B. Key, E.D. Strozie r, S.B. Galloway 199 8 Investigation of unusual mortalities of bottlenose dolphins along the mid Texas coastal bay ecosystem during 1992. NOAA Technical Report. NMFS 147. Colman, J.R. and J.S. Ramsdell. 2003. The type B (PbTx 3) brevetoxin advers ely affects development, cardiovascular function, and survival in medaka ( Oryzias latipes ) embryos. Environmental Health Perspectives 111: 1920 1925. Cornell Lab of Ornithology. 2003. All about birds: online bird guide. Available online at: http://www.birds.cornell.edu/AllAboutBirds/BirdGuide/ Last accessed October 2007. Coulson, J.C., G.R. Potts, I.R. Deans, and S.M. Fraser. 1968. Exceptional mortality of shags and other se abirds caused by paralytic shellfish poison. British Birds 61(9): 381 407.
56 Cummins, J.M., A.C. Jones and A.A. Stevens. 1971. Occurrence of toxic bivalve molluscs during a Gymnodinium breve red tide. Transactions of the American Fisheries Society 10 0(1): 112 116. Duffus, J.H. and H.G.J. Worth. 2001. The science of chemical safety and essential toxicology: an educational resource. International Union of Pure and Applied Chemistry Electronic Resource. Available online at: www.iupac.org/publications/cd/essential_toxicology/index.html Last accessed October 2007. Dunning, J.B. 1993. CRC Handbook of Avian Body Masses. CRC Press, Inc. Boca Raton, FL. 371 p p. Elphick, C., J.B. Dunning, Jr. and D.A. Sibley. 2001. The Sibley Guide to Bird Life and Behavior. Alfred A. Knopf, New York. 608 pp. Emslie, S.D., W.D. Allmon, F.J. Rich, J.H. Wrenn, and S.D. de France. 1996. Integrated taphonomy of an avian de ath assemblage in marine sediments from the late Pliocene of Florida. Palaeogeography, Palaeoclimatology, and Palaeoecology 124: 107 136. Ernst, E. June 16, 2003. Treatment aids red tide stricken dogs. Sarasota Herald Tribune, pp. 1B 2B. Flemi ng, L.E. and D.G. Baden. 1998. Neurotoxic Shellfish Poisoning: Public Health and Human Effects. Proceedings, Texas Neurotoxic Shellfish Poisoning Conference, Corpus Christi, TX. 6 pp. Flewelling, L., J.P. Naar, J.P. Abbot, D.G. Baden, N.B. Barros, G. D. Bossart, M.D. Bottein, D.G. Hammond, E.M. Haubold, C.A. Heil, M.S. Henry, H.M. Jacocks, T.A. Leighfield, R.H. Pierce, T.D. Pitchford, S.A. Rommel, P.S. Scott, K.A. Steidinger, E.W. Truby, F.M. Van Dolah and J.H. Landsberg. 2005. Brevetoxicosis: r ed t ides and marine mammal mortalities. Nature 435(7043): 755. Florida Fish and Wildlife Conservation Commission (F W RI). 2006. Species Account Atlantic thread herring ( Opisthonema oglinum ). St. Petersburg, FL. 7 pp. Florida Fish and Wildlife Conserva tion Commission (F W RI). 2006. Species Account Pinfish ( Lagodon rhomboides ). St. Petersburg, FL. 9 pp. Florida Fish and Wildlife Conservation Commission (F W RI). 2006. Species Account Spanish sardine ( Sardinella aurita ). St. Petersburg, FL. 4 pp
57 Forrester, D.J., J.M. Gaskin, F.H. White, N.P. Thompson, J.A. Quick, G.E. Henderson, J.C. Woodard, and W.D. Robertson. 1977. An epizootic of waterfowl associated with a red tide episode in Florida. Journal of Wildlife Diseases 13: 160 167. Geraci, J.R., D.M. Anderson, R.J. Timperi, D.J. Staubin, G.A. Early, J.H. Prescott, and C.A. Mayo. 1989. Humpback whales ( Megaptera novaeangliae ) fatally poisoned by dinoflagellate toxin. Canadian Journal of Fisheries and Aquatic Science 44: 1895 1898. Gochfe ld, M. and J. Burger. 1980. Opportunistic scavenging by shorebirds: feeding behavior and aggression. Journal of Field Ornithology 51 (4): 373 375. Gore, R.H. 1992. The Gulf of Mexico. Pineapple Press, Inc. Sarasota Florida. 384 pp. Gunter, G., F.G.W. Smith and R.H. Williams. 1947. Mass mortality of marine animals on the lower west coast of Florida, November 1946 to January 1947. Science 105(273): 256 257. Gunter, G., R.H. Williams, C.C. Davis and R.G.W. Smith. 1948. Catastrophic mass mor tality of marine animals and coincident phytoplankton bloom on the west coast of Florida, November 1946 to August 1947. Ecological Monographs 18: 308 324. Gunter, G. 1958. Feeding behavior of brown and white pelicans on the Gulf coast of the United S tates. Louisiana Academy of Sciences 21: 34 39. Hallegraeff G.M. 1993. A review of harmful algal blooms and their apparent global increase. Phycologia 32(2): 79 99. Hockey, P.A.R. and J. Cooper. 1980. Paralytic shellfish poisoning a controll ing factor in black o ystercatcher populations? Ostrich 51: 188 190. International Labor Organization (ILO) United Nations Environmental Program (UNEP) World Health Organization (WHO). 1984. Aquatic (marine and freshwater) biotoxins. Environmental Health Criteria, vol. 37. WHO ILO, Geneva. Kennedy, C., L.S. Schulman, D.G. Baden, P. Walsh. 1992. Toxicokinetics of brevetoxin PbTx 3 in the gulf toadfish, Opsanus beta following intravenous administration. Aquatic Toxins 22: 3 14. Kimball, K .D. and S.A. Levin. 1985. Limitations of laboratory bioassays: the need for ecosystem level testing. Bioscience 35: 165 171.
58 Kirkpatrick, B., L.E. Fleming, D. Squicciarini, L.C. Backer, R. Clark, W. Abraham, J. Benson, Y.S. Cheng, D. Johnson, R. Pie rce, J. Zaias, G.D. Bossart and D.G. Baden. 2004. Literature review of Florida red tide: implications for human he alth effects. Harmful Algae 3 : 99 115. Kress, S.W. 2000. National Audubon Society Birders Handbook. Dorling Kindersley, New York. 176 pp. Kreuder, C., J.A.K. Mazet, G.D. Bossart, T. Carpenter, M. Holyoak, M. Elie and S. Wright. 2002. Clinicopathologic features of suspected brevetoxicosis in double crested cormorants ( Phalacrocorax auritus ) along the Florida G ulf coast. Journal o f Zoo and Wildlife Medicine 33: 8 15. Landsberg, J.H., Tester, P., Pierce, R., Shea, D., Van Dolah, F., Sosa, E., Henry, M., Fournie, J., Flewelling, L., Varnam, S. and Leighfield, T. 2000. ECOHAB FLORIDA: Fate and effects of brevetoxins in selected b iota, water and sediments along the West Florida Shelf, USA. Presented at The Symposium on Harmful Marine Algae in the US, Woods Hole, MA. December 2000. Landsberg, J.H. 2002. The effects of harmful algal blooms on aquatic organisms. Reviews in Fishe ries Science 10(2): 113 390. Lefebvre, K.A., C.L. Powell, M. Busman, G.J. Doucette, P.D.R. Moeller, J.B. Silver, P.E. Miller, M.P. Hughes, S. Singaram, M.W. Silver, and R.S. Tjeerdema. 1999. Detection of domoic acid in northern anchovies and Californ ia sea lions associated with an unusual mortality event. Natural Toxins 7: 85 92. Lefebvre, K.A., S.L. Dovel, and M.W. Silver. 2001. Tissue distribution and neurotoxic effects of domoic acid in a prominent vector species, the northern anchovy Engraulis m ordax. Marine Biology 138: 693 700. Lefebvre, K.A., M.W. Silver, S.L. Coale, and R.S. Tjeerdema. 2002. Domoic acid in planktivorous fish in relation to toxic Pseudo nitzschia cell densities. Marine Biology 140: 625 631. Lefebvre, K.A., S. Bargu, T. Kieckhefer, and M.W. Silver. 2002b. From sanddabs to blue whales: the pervasiveness of domoic acid. Toxicon 40: 971 977. Montevecchi, W.A., 1993. Birds as indicators of change in marine prey stocks. P p. 217 266 In: R.W. Furness and J.J.D. Gree nwood ( Eds. ). Birds as Monitors of Environmental Change. Chapman and Hall, London Montevecchi, W.A., 2001. Seabirds as indicators of ocean pollution. In: J. Steele, S. Thorpe, K. Turekin ( Eds. ). Encyclopedia of Ocean Sciences. Academic Press, Londo n.
59 Naar, J., A. Bourdelais, C. Tomas, J. Kubanek, P.L. Whitney, L. Flewelling, K. Steidinger, J. Lancaster, and D.G. Baden. 2002. A competitive ELISA to detect brevetoxins from Karenia brevis (formerly Gymnodinium breve ) in seawater, shellfish and mammal ian body fluid. Environmental Health Perspectives 110(2): 179 185. Nakanishi, K. 1985. The chemistry of brevetoxins: a review. Toxicon 23(3): 473 479. Nisbet, I.C.T. 1983. Paralytic shellfish poisoning: Effects on breeding terns. Condor 85: 3 38 345. Owre, O.T. 1990. Coastal and Marine Birds. In: Continental Shelf Associates, Inc. (Ed.). Synthesis of a vailable biological, g eological, chemical, s ocioeconomic, and c ultural r esource i nformation for the South Florida a rea. U.S. Department of the Interior, Minerals Management Service, Atlantic OCS Region, pp. 303 351. Pierce, R.H. 1986. Red tide ( Ptychodiscus brevis ) toxin aerosols: A review. Toxicon 24(10): 955 965. Poli, M.A., S.M. Musser, R.W. Dickey, P.P. Eilers, S. Hall. 2000. Neu rotoxic shellfish poisoning and brevetoxin metabolites: a case study from Florida. Toxicon 38: 981 993. Poli, M.A., C.B. Templeton, W.L. Thompson and J.F. Hewetson. 1990. Distribution and elimination of brevetoxin PbTx 3 in rats. Toxicon 28(8): 903 910. Quick, J.A. and G.E. Henderson. 1974. Effects of Gymnodinium breve red tide on fishes and birds: preliminary report on behavior, anatomy, hematology and histopathology. Pp. 85 113. In: Proceedings o f the Gulf Coast regional symposium on d iseases of aquatic animals. Louisiana State University, Baton Rouge, LA. Radwan, F.F.Y, Z. Wang and J.S. Ramsdell. 2005. Identification of a rapid detoxification mechanism for brevetoxin in rats. Toxicological Sciences 85: 839 846. Rand, A.L. 195 7. Sanderlings eat fishermens bait m innows. The Wilson Bulletin 69 (2): 186 187. Ray, S.M. and D.V. Aldrich. 1965. Gymnodinium breve : i nduction of shellfish p oisoning in c hicks. Science 148(3678): 1748 1749. Roberts, B.S.A., G.E. Henderson and R.A. Medlyn. 1979. The effect of Gymnodinium breve toxin(s) on selected mollusks and crustaceans. Pp. 419 424. In : D.L. Taylor and H.H. Seliger (Eds.). Toxic Dinoflagellate Blooms. Elsevier, North Holland.
60 Runde, D.E. 1991. Trends in Wading Bird Nesting Populations in Florida, 1976 1978 and 1986 1989. Final Performance Report. Florida Game and Fresh Water Fish Commission. Tallahassee, FL. 90 pp. Schramm, H.L., Jr., M.W. Coll o py and E.A. Okrah. 1987. Potential problems of bird predation for fish culture in Florida. Progressive Fish Culturist 49: 44 49. Scholin, C.A., F. Gulland, G.J. Doucette, S. Benson, M. Busman, F.P. Chavez, J. Cordaro, R. DeLong, A. De Vogelaere, J. Harvey, M. Haulena, K. Lefebvre, T. Lipscomb, S. Loscuttoff, L.J. Lo wenstine, R. Marin III, P.E. Miller, W.A. McLellan, P.D.R. Moeller, C.L. Powell, T. Rowles, P. Silvagni, M. Silver, T. Spraker, V. Trainer and F.M. Van Dolah. 2000. Mortality of sea lions along the central California coast linked to a toxic diatom bloom. Nature 403: 80 84. Schreiber, E.A. and J. Burger (Eds.). 2002. Biology of marine birds. CRC Press LLC. Boca Raton, FL. 722 pp. Schreiber R.W. and R. B Clapp. 1987. Pelecaniform feeding ecology. Pp. 173 88. In: J. P. Croxall ( Ed. ) Seabirds : f eeding ecology and role in marine ecosystems. Cambridge Univ. Press, Cambridge. Shumway, S.E., S.M. Allen and P.D. Boersma. 2003. Marine birds and harmful algal blooms: sporadic victims or under reported events? Harmful Algae 2: 1 17. Sierra Bel trn, A., M. Palafox Uribe, J. Grajales Montiel, A. Cruz Villacorta, and J.L. Ochoa. 1997. Sea bird mortality at Cabo San Lucas, Mexico: evidence that toxic diatom blooms are spreading. Toxicon 35: 447 453. Simon J L and D .M. D auer 1972 A quant itative evaluation of red tide induced mass mortalities of benthic invertebrates. Environ mental Lett ers 3: 229 234 Smayda T.J. 2002. Adaptive ecology, g rowth strategies and the g lobal bloom e xpansion of d inoflagellates. Journal of Oceanography 58: 281 294 Smith, G.B. 1975. The 1971 red tide and its impact on certain reef communities in the eastern Gulf of Mexico. Environmental Letters 9(2): 141 152. Steidinger, K.A. and K. Haddad. 1981. Biologic and h ydrographic a spects of r ed t ides. Biosci ence 31(11): 814 819. Steidinger, K.A. and R.M. Ingle. 1972. Observations on the 1971 red tide in Tampa Bay, Florida. Environmental Letters 3: 271 278.
61 Steidingder, K.A. and G.A. Vargo. 1988. Marine dinoflagellate blooms: dynamics and i mpacts. Pp 373 401. In: C.C. Lembi and J.R. Waaland, (E ds. ). Algae and Human Affairs, Cambridge Press, New York. Steidinger, K.A., G.A. Vargo, P.A. Tester, and C.R. Tomas. 1998. Bloom dynamics and physiology of Gymnodinium breve with emphasis on the Gulf of Mexico. P p 133 153. In: D.M. Anderson, A.D. Ce mbella and G.M. Hallengraff (Eds.). Physiological Ecology of Harmful Algal Blooms. NATO ASI Series G: Ecological Science, Volume 41. Springer Verlag, New York. Tester P.A., J.T. Turner and D. Shea 2000. Vectorial transport of toxins from the dinoflagellate Gymnodinium breve through copepods to fish. Journal of Plankton Research 22(1): 47 61. Turgeon, D.D., K.G. Sellner, D. Scavia, and D. Anderson. 1998. Status of U.S. harmful algal blooms: Progress towards a national program. U.S. Department of Commerce, NOAA, October 1998. 16 pp. United States Fish and Wildlife Service, Division of Migratory Bird Management. Double crested Cormorants. Available online at: http://www.fws.gov/migratorybirds/issues/cormorant/cormorant.html Last accessed October 2007. Van Dolah F.M. 2000. Marine algal toxins: origins, health effects and their increased occurrence. E nvironmental Health Perspectives 108: 133 141. Vargo, G.A., K.L. Carder, W. Gregg, E. Shanley, C. Heil, K.A. Steidinger and K.D. Haddad. 1987. The potential contribution of primary production by red tides to the West Florida Shelf ecosystem. Limnology and Oceanography 32: 762 767. Vargo, G.A. 1999. Coastal Phytoplankton blooms: What are they telling us? ASB Bulletin Coastal Ecology Symposium. 46(4): 286 309. Vargo, G.A., K. Atwood, M. van Deventer and R. Harris. 2006. Beached bird surveys on Shell Key, Pinellas County, Florida. Florida Field Naturalist 34(1): 21 27. Vidal Hernandez, L. and S. Nesbitt. 2002. Seabirds. Pp. 108 109 In: T.A. Okey and B. Mahmoudi ( Eds. ) An ecosystem model of the West Florida Shelf for use in fisheries m anagemen t and ecological research. Volume II. Florida Marine Research Institute, St. Petersburg, FL. Williams, E.H., L. Bunkley Williams, and I. Lopez Irizarry. 1992. Die off of Brown Pelicans in Puerto Rico and the United States Virgin Islands. Sc ience 46: 1106 1108.
62 Wires, L.R., F.J. Cuthbert, D.R. Trexel and A.R. Joshi. 2001. Status of the Double c rested Cormorant ( Phalacrocorax auritus ) in North America. Final Report to USFWS. 359 pp. Woofter, R.T., K. Brendtro and J.S. Ramsdell. 2005. Uptake and Elimination of brevetoxin in blood of striped mullet ( Mugil cephalus ) after aqueous exposure to Karenia brevis Environmental Health Perspectives 113(1): 11 16. Work, T.M., B. Barr, A.M. Beale, L. Fritz, M.A. Quilliam, and J.L.C. Wright. 1993 Epidemiology of domoic acid poisoning in brown pelicans ( Pelecanus occidentalis ) and Brandt's cormorants ( Phalacrocorax penicillatus ) in California. Journal of Zoo and Wildlife Medicine 24: 54 62. Work, T.M. 2000. Avian necropsy manual for biologists in remote refuges. United States Geological Survey National Wildlife Health Center, Hawaii Field Station. Available online at: http://www.nwhc.usgs .gov/publications/necropsy_manuals/Avian_Necropsy_Ma nual English.pdf [Last accessed October 2007].
64 APPENDIX I : Mean weights for all species for which more than one individual was sampled during this study. Healthy adult mea n ranges are listed for comparison. Sample Size Observed Weights (g) Adult Mean:Range Weight (g) Species Mean Std Dev Range Double crested Cormorant 18 1160 249.8 805 1560 1800 3000 Brown Pelican 10 2140 458.1 1420 2950 3200 3700 L aughing Gull* 11 230 58.8 155 356 289 327 Sanderling 11 52 6.2 40 63 40 100 Sandwich Tern 5 125 22.3 93 148 140 300 Great Blue Heron** 4 1771 1036.3 284 2686 2100 2500 Northern Gannet 4 1713 207.4 1517 1932 2930 3070 Osprey 4 1101 196.1 852 1307 1400 2000 Royal Tern 3 416 95.3 306 475 320 500 Source: Schreiber and Burger, 2002. Source: Cornell Lab of Ornithology, 2003. *Some individuals were partially decomposed, therefore listed weights m ay not provide accurate indication of body condition at time of death. **Includes one nestling.
65 APPENDIX II : Brevetoxin findings for shorebirds collected during reported mortality event in the study region from August to October, 2005. Species Date of Death Location Recovered Liver Stomach Intestine Sanderling 8/12/2005 N. Captiva Beach 202.39 66.40 187.04 Sanderling 8/26/2005 Dunedin 48.75 28.81 29.48 Ruddy Turnstone 8/29/2005 Clearwater Beach 46.26 58.35 44.56 Sanderling 9/4/2005 Sand Key 125.32 167.91 196.12 Sanderling 9/4/2005 Sand Key 38.84 19.01 31.05 Sanderling 9/5/2005 Sand Key 107.30 49.46 57.27 Willet 9/6/2005 Sand Key 254.64 238.60 91.76 Ruddy Turnstone 9/10/2005 Dunedin 25.73 80.99 117.38 Sanderling 9/22/2005 Fort DeSoto 201 .90 119.69 129.05 Sanderling 9/22/2005 Fort DeSoto 74.28 105.93 47.97 Black bellied Plover 9/24/2005 Isla del Sol 53.73 1267.42 122.88 Sanderling 10/12/2005 Fort DeSoto 115.76 185.31 126.18 Sanderling 10/14/2005 Indian Shores 135.19 239.18 82.49 Mea n 110.01 202.08 97.17 Std Dev ( ) 72.59 328.60 55.36
66 APPENDIX III : Results for all double crested cormorants ( P. auritus ) collected during this study, and indication of brevetoxicosis diagnosis by attending rehabilitation facility. Location Fou nd Date of Death Red Tide? Liver Stomach Contents Intestinal Contents Gall Bladder DCCO1 Sanibel 1/8/2005 Yes
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001935410
007 cr mnu|||uuuuu
008 080423s2007 flua sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002291
Van Deventer, Michelle.
Brevetoxins in marine birds :
b evidence of trophic transfer and the role of prey fish as toxin vector
h [electronic resource] /
by Michelle van Deventer.
[Tampa, Fla.] :
University of South Florida,
ABSTRACT: Harmful algal blooms (HABs) of the brevetoxin-producing dinoflagellate Karenia brevis occur periodically along the central west coast of Florida. Mass mortalities of marine birds have long been associated with these blooms, yet there is little data documenting the accumulation of brevetoxins in the tissues of birds and their prey items. An intense HAB event impacted the region from Tampa Bay to Charlotte Harbor during most of 2005. More than one hundred marine birds, representing twenty three species, were collected during this bloom. All birds sampled were found dead or had died within 24 hours of admittance to local wildlife rehabilitation centers. In order to determine if fish were vectors for brevetoxin ingestion, the stomach contents of all birds were examined and any recovered fish were identified to the extent possible.The gastrointestinal tissues and contents from all avian samples were analyzed for brevetoxin levels, with results ranging from < LD to 9988.62 ng PbTx per gram tissue. Small planktonivorous fish such as thread herring, sardines and anchovies that largely comprise the diet of affected piscivorous birds were also collected and analyzed for brevetoxin content, with results ranging from < LD to 5839.90 ng PbTx per gram tissue. The highest levels of brevetoxins were generally detected in the viscera of fish, with relatively low levels detected in the muscle tissues. These results indicate that piscivorous marine birds, including double-crested cormorants, brown pelicans, terns and gulls, are exposed to a range of brevetoxin levels in their diet during Karenia brevis blooms. Ingestion appears to be the primary route of exposure, and brevetoxin-contaminated fish were confirmed in the stomachs of several birds.Shorebirds and gulls may also be exposed to brevetoxins via scavenging of red tide-killed fish deposited on beaches during blooms. Samples from scavenged fish were found to have brevetoxin levels ranging from 31 to 95,753 ng PbTx per gram tissue.
Thesis (M.S.)--University of South Florida, 2007.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 66 pages.
Advisor: Gabriel A. Vargo, Ph.D.
x Marine Science
t USF Electronic Theses and Dissertations.