Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The risk of non-target species poisoning from brodifacoum used to eradicate rats from Langara Island,… Howald, Gregory Robert 1997

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

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
831-ubc_1997-0545.pdf [ 6.41MB ]
Metadata
JSON: 831-1.0087971.json
JSON-LD: 831-1.0087971-ld.json
RDF/XML (Pretty): 831-1.0087971-rdf.xml
RDF/JSON: 831-1.0087971-rdf.json
Turtle: 831-1.0087971-turtle.txt
N-Triples: 831-1.0087971-rdf-ntriples.txt
Original Record: 831-1.0087971-source.json
Full Text
831-1.0087971-fulltext.txt
Citation
831-1.0087971.ris

Full Text

T H E RISK O F N O N - T A R G E T S P E C I E S P O I S O N I N G F R O M B R O D I F A C O U M U S E D T O E R A D I C A T E R A T S F R O M L A N G A R A ISLAND, BRITISH C O L U M B I A , C A N A D A by G R E G O R Y R O B E R T H O W A L D B.Sc. (Agr.)(Hons.), University of British Columbia, Vancouver, 1993. A THES IS IN PARTIAL FULF ILLMENT O F T H E R E Q U I R E M E N T S F O R THE D E G R E E OF M A S T E R O F S C I E N C E in T H E F A C U L T Y O F G R A D U A T E STUDIES Department of Animal Sc ience We accept this thesis as conforming to the required standard The University of British Columbia August, 1997 © Gregory Robert Howald In presenting this thesis in partial fulfilment of the ; requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for-extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by. his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of AoiflAL ^ c ICpOL The University of British Columbia Vancouver, Canada Date OCTO&E& 6 /^ DE-6 (2/88) ^ Abstract In 1995, the Canad ian Wildlife Serv ice attempted to eradicate introduced Norway rats (Rattus norvegicus) from Langara Island and adjacent Cox and Lucy Islands with the application of the second generation anticoagulant rodenticide brodi facoum. However, anticoagulant rodenticides are non-species specif ic pest ic ides and pose a poisoning risk to non-target spec ies. This thesis addresses the short term poisoning impacts to non-target spec ies from brodifacoum used to eradicate the rats from Langara and Lucy Island. In 1994, during testing of the baiting protocol on Lucy Island, the native dusky shrew (Sorex monticolus elassodon) population s ize fell from an est imated 25 unique shrews/ha before the baiting to four unique shrews/ha after the baiting. This prompted a monitoring program in three regions on Langara Island in 1995. Whi le shrews were attracted to bait in stations, the decl ine in their population was non-significant. Shrews in breeding condition were at greater risk of poisoning likely due to their ability to range widely. Shrews entered and chewed on bait blocks in up to 8 0 % of bait stations. The risk of secondary poisoning to avian scavengers from poisoned toxic rat ca r casses was investigated. In 1994, common ravens (Corvus corax) were identified as the most significant scavenger of rat carcasses . In 1995, two of 15 radio-collared Norway rats poisoned with brodifacoum died above ground and one was scavenged . Who le body brodifacoum residues from other rats found dead above ground ranged from 2.40-16.51 mg/kg. Between 1995 and 1996, 20 raven ii remains were found or reported. In 1995, 13 raven livers tested positive for brodi facoum. Ravens were secondari ly poisoned from scavenging rat ca r casses and primarily from raiding bait stations. Brodi facoum was detected in Northwestern crows [Corvus caurinus) 9 months after the cessat ion of baiting on Lucy Island in 1994, but before the baiting on Langara Island. Brodi facoum residues were detected in the p lasma of 15% of bald eagles [Haliaeetus leucocephalus) sampled (0.037-1.74 ppm). The invertebrates as a source of brodifacoum to non-target spec ies was investigated. Snai ls (Vespericola sp. and Haplotrema sp.) and banana s lugs (Ariolimax sp.) were common and abundant invertebrates found feeding on bait in stations. The blue coloured bait could be seen through the translucent bodies and the mol luscs tested positive for brodifacoum. Carr ion insects readily consumed rat ca r casses containing brodifacoum. Blowfly larva {Calliphora sp.) tested positive for brodifacoum residues. The invertebrates found feeding on the bait and carrion insects were a secondary and tertiary poisoning risk to non-target spec ies such as the song sparrow (Melospiza melodia). Table of Contents Abstract ii Tab le of Contents iv List of Tab les vi List of F igures ix List of Append i ces xii Abbreviat ions xiii Acknowledgements xiv Dedicat ion xv Chapter 1. General Introduction 1 1.1 Object ives of Research 4 1.2 Study Areas 5 1.3 Bait ing Protocol 9 Chapter 2. Hemostasis and Mode of Action of the Anticoagulant Rodenticides: the Physiological Basis for Non-Target Species Poisoning Concern 13 2.1 Introduction 13 2.2 Hemostas is 13 2.3 Vitamin K 1 and the Mode of Act ion of the Ant icoagulants 16 Chapter 3. The Short Term Impacts of Brodifacoum Baiting on the Native Small Mammals 20 3.1 Introduction 20 3.2 Materials and Methods 22 3.3 Resul ts 27 3.4 Discuss ion 34 Chapter 4. An Evaluation of the Secondary Poisoning Hazard to Avian Wildlife 38 4.1 Introduction 38 4.2 Materials and Methods 41 4.3 Resul ts 49 4.4 Discuss ion 66 iv Chapter 5. The Uptake of Brodifacoum by Invertebrates Feeding on Bait Containing the Rodenticide Brodifacoum or Norway Rat carcasses Poisoned with Brodifacoum 89 5.1 Introduction 89 5.2 Materials and Methods 90 5.3 Results 97 5.4 D iscuss ion 108 Chapter 6. Conclusions and Recommendations 112 6.1 Conc lus ions 112 6.2 Recommendat ions 114 References 120 v List of Tables Table 3.1. Abundance of deer mice (Peromyscus maniculatus) /100 Trap Nights on G r aham Island, before and after the intensive baiting period on Langara Island, 1995 27 Table 3.2. Dusky shrew (Sorex monticolus) trap success (# shrews caught/100 Trap Nights) and population estimate before and after brodifacoum bait appl ication, Lucy Island, 1994/1995 28 Table 3.3. Proportion of dusky shrews (Sorex monticolus) in breeding condition (testes scrotal, nipples large) pre and post baiting on Lucy Island, 1994/1995 (n in brackets) 29 Table 3.4. Dusky shrew (Sorex monticolus) trap success (# caught/100 Trap Nights) by trap sess ion and region on Langara Island and G raham Island, 1995...30 Table 3.5. Proportion of dusky shrews (Sorex monticolus) in breeding condition before and after intensive baiting, Langara Island, 1995 (mean of 2 grids per region; sample s ize in brackets) 32 Table 3.6. Number of captures, weight, maximum distance travelled and 9 0 % M C P range s ize for dusky shrews (Sorex monticolus) caught five or more t imes, Langara Island, 1995 33 Table 4.1. Ba ld eagle (Haliaeetus leucocephalus) trapping success using the floating fish set, Langara Island, 1995 (n=148 sets) 46 Table 4.2. Bald eagle (Haliaeetus leucocephalus) trapping results, Langara Island 1995 46 Table 4.3. Quantitative recovery of brodifacoum from Norway rat (Rattus norvegicus) liver fortified on Langara Island in August, 1995 48 Table 4.4. Ca r cas s locations of brodifacoum poisoned radio-tagged Norway rats (Rattus norvegicus), Langara Island, 1995 50 Table 4.5. Interval between start of poisoning and detected death of radio-tagged Norway rats (Rattus norvegicus), Langara Island, 1995 50 Table 4.6. Brodi facoum residue concentrations (mg/kg) in Norway rats (Rattus norvegicus) found dead above ground, Langara Island, 1995; (means, 9 5 % conf idence limits in brackets; means that do not share the same lowercase letter were significantly different at p<0.05) 53 vi Table 4.7. Brodi facoum residues (mg) in Norway rats (Rattus norvegicus) found dead above ground, Langara Island, 1995; (mean±s.e., range in brackets). 54 Table 4.8. Frequency of avian scavengers of unpoisoned Norway rat (Rattus norvegicus) carcasses , Langara Island, 1994 54 Table 4.9. Identified scavengers of unpoisoned Norway rat (Rattus norvegicus) carcasses , Langara Island, 1994 55 Table 4.10. Brodi facoum residues, gut contents, and sites of hemorrhage to common ravens (Corvus corax) found dead, Langara Island, 1995 57 Table 4.11. C ommon raven (Corvus corax) liver brodifacoum residue levels (mg/kg), Langara Island, 1995 (geometric mean; 9 5 % conf idence interval in brackets) 58 Table 4.12. Pr imary sites of hemorrhage in common ravens (Corvus corax) found dead, Langara Island, 1995 58 Table 4.13. Food remains in the gizzards and intestines of 13 common ravens (Corvus corax) found dead, Langara Island, 1995 59 Table 4.14. Inactive common raven (Corvus corax) nests, Langara Island, 1996 60 Table 4.15. Dates, location and condition of common ravens (Corvus corax) found or reported dead, Langara Island, 1996 62 Table 4.16. Brodi facoum residue levels in livers of Northwestern Crows (Corvus caurinus), Langara Island, 1995 63 Table 4.17. Bald eagle (Haliaeetus leucocephalus) p lasma brodifacoum res idues and prothrombin t imes, Langara Island, 1995 64 Table 4.18. Bald eagle (Haliaeetus leucocephalus)contro\ prothrombin t imes 65 Table. 4.19. Spec ies and L D 5 0 values used for calculating the value that offers 9 5 % bird spec ies protection with 95% and 50% confidence limits 77 Table 5.1. Brodi facoum residue recovery from fortified samples, 1994 97 vii Table 5.2. Brodi facoum residue recovery rate from fortified samples prepared on Langara Island, 1995 98 Table 5.3. Brodi facoum residues in invertebrates found in bait stations, Langara Island, 1995 102 Table 5.4. Brodi facoum residues in invertebrates found in bait stations, Lucy Island, 1994 103 Table 5.5. Number of and brodifacoum residues in carrion insects col lected from brodifacoum poisoned Norway rats and control Norway rats, into 9 5 % ethanol, Langara Island, 1994 105 Table 5.6. Insects col lected and preserved in 9 5 % ethanol with no brodifacoum residues detected, Langara Island, 1995 106 Table 5.7. Brodi facoum residues in carrion insects col lected from brodifacoum poisoned Norway rat carcasses , col lected fresh, Langara Island, 1994 107 Table 5.8. Brodi facoum residues in blowfly larva col lected from brodifacoum po isoned Norway rat carcasses , Langara Island, 1995 107 Table A-1. Liver brodifacoum residue in Japanese Quai l after a single oral dose of brodifacoum (mean ±s.e., n=2 for each time/dose group) 144 Table B-1. Brodi facoum residues detected in Song Sparrows, Langara Island, 1995 152 Table B-2. Morphological measurements from Song Sparrows col lected on Langara Island, 1995 152 viii List of Figures Figure 1.1. Chemica l structure of brodifacoum. The more common and familiar anticoagulant warfarin is presented for compar ison purposes. Note the 4-hydroxycoumarin ring system common in both brodifacoum and warfarin 3 Figure 1.2. Locat ion of Langara, Lucy and Cox Islands on the British Co lumbia Coast , C a n a d a 7 Figure 1.3. P lace names on Langara Island, 1995 8 Figure 1.4. Bait station des ign used on Lucy, Langara and Cox Islands. The stations were made of orange coloured P V C corrugated drainage pipe. The lid was removable for p lacement of bait and checking of stations. The lid snapped into p lace to protect the bait from the weather and most non-target spec ies . Each station was labelled with a unique identification number. The stations were staked to the ground with 2-60 cm long wires bent into a horseshoe shape and placed on either end.(adapted from Taylor 1993) 10 Figure 1.5. Bait station Layout on Langara, Lucy and Cox Islands, 1995. Each dot represents 1 bait station. Total number of stations: 3848 12 Figure 2.1. A diagrammatic representation of the coagulat ion cascade . The grey arrows refer to the intrinsic pathway. The roman numerals represent the clotting factors. The lower case a, represents the activated form of that clotting factor. Dashed l ines represent positive feedback loop enhancing the coagulat ion cascade , (adapted from Sturkie 1986 and Rapaport 1987).... 15 Figure 2.2. The Vitamin K Metabol ic Activities in Rat Liver M icrosomes. (Black arrows indicate where the anticoagulants inhibit the cycl ing of vitamin K) Step A is a carboxy lase-epox idase enzyme not inhibited by coumarins. Step B, is the epox ide reductase enzyme which is inhibited by the anticoagulants. This leads to rapid depletion of Vitamin K stores and build up of Vitamin K epoxide. Step C is a lso a reductase enzyme inhibited by anticoagulants, however, an alternative pathway exists al lowing this reaction to proceed. (Adapted from Mount et al. 1982; Suttie 1980) 17 Figure 3.1. Location of small mammal trapping grids, Langara and Lucy Island, 1995 23 Figure 3.2. The number of unique dusky shrews (Sorex monticolus) captured on Lucy Island, 1994 and 1995. The vertical arrows indicate when the brodifacoum baiting for introduced rats began. The bait was removed in mid August, 1994, but was reapplied in July 1995 28 ix Figure 3.3. Mean unique number of shrews captured on Langara Island and on the control grids on G raham Island, 1995. Vertical arrow indicates when brodi facoum baiting for introduced rats began on Langara Island 29 Figure 3.4. Unique number of dusky shrews captured on each study grid in the 3 regions on Langara Island, 1995. Vertical arrows indicate when brodifacoum baiting began. Longworth traps were used on all grids except where denoted by a * when smal l Shermann traps were used 31 Figure 3.5. The proportion of dusky shrews (Sorex monticolus) in breeding condition correlated to the post baiting shrew population estimate as a proportion of the pre- baiting population estimate (r = 0.64) 36 Figure 4.1. Locat ions of all ca rcasses of common raven [Corvus corax) and Norway rats (Rattus norvegicus), together with sampl ing locations of bald eag les (Haliaeetus leucocephalus) and Northwestern crows (Corvus caurinus) with positive detection of brodifacoum residue, 1995 52 Figure 5.1. Carr ion Insect Trap used to capture and hold insects attracted to rat ca r casses on Langara Island, 1994. Note the funnel over the opening of the tube on the left hand side, and the 2.5 cm diameter P V C tubing exiting the tube into the 500 ml g lass jar for holding insects. The g lass jar has small air holes in the removable lid. Insects were col lected at regular intervals from the g lass jar and frozen. The black triangles were the ventilation holes and were covered with noseum netting 93 Figure 5.2. Proportion of bait stations with snai ls and banana s lugs over the course of the intensive baiting period, Langara Island, 1995 (number of bait stations in brackets) 100 Figure 5.3. Proportion of bait stations with the terrestrial snai ls, Vespericola sp. and Haplotrema sp., over the course of the intensive baiting period (number of bait stations in brackets) 101 Figure A-1. Effect of time on brodifacoum p lasma residue concentration (ppm) after a single oral dose of brodifacoum at 0.35 and 0.7 mg/kg (means that do not share the s ame letter were significant at P<0.05)(n=4 at each time point) 142 Figure A-2. Effect of dose on brodifacoum residue concentration (ppm) over 10 days (* significant at P<0.05)(n=2 at each dose level) 143 Figure A-3. Effect of time on the prothrombin time ratio (PTR) of J apanese Quai l after a single oral dose of brodifacoum (* significant at P<0.05)(n=9 at each time point) 145 x Figure A-4. Effect of dose of the prothrombin time ratio (PTR) of J apanese quail after a single oral dose of brodifacoum (means that do not share the same letter were significantly different at P<0.05) 147 xi List of Appendices Appendix Table 4-1. Dates, locations, weight, age and sex of Norway rats (Rattus norvegicus) found dead above ground, Langara Island, 1995 85 Appendix Table 4-2. Dates, locations, and selected morphological measurements of bald eag les (Haliaeetus leucocephalus), Langara Island, 1995 86 Appendix Table 4-3. Dates, locations and selected morphological measurements of Northwestern crows (Corvus caurinus) col lected on Langara and Lucy Island, 1995 87 Appendix A. The Detection of Exposure to Brodi facoum in J apanese Quai l through P l a sma Brodi facoum Res idue Analys is and Prothrombin T ime Evaluat ion 134 Appendix B. Brodi facoum Exposure in the Song Sparrow 151 Appendix C. Potential Sub-Lethal and Long Term Effects of Brodi facoum Exposure 154 Appendix D. Environmental Aspec ts of Brodi facoum - Transport, Distribution and Transformation 157 xii Abbreviations A N O V A analys is of var iance ppm parts per million C W S Canad ian Wildlife Serv ice PT prothrombin time dbh diameter at breast height P W R C Pacif ic Wildlife Research Centre GIT gastro-intestinal tract s standard deviation ha hectare S A S trademark, S A S Institute Inc. H A G height above ground s.e. standard error ND none detected ug micro gram N W R C National Wildlife Research Centre xiii Acknowledgements I would like to thank my immediate thesis supervisors, Dr. K im Cheng at U B C , Dr. John Elliott at P W R C and Dr. Pierre Mineau from N W R C , for their support, gu idance and ass is tance both in the field and lab. I am also grateful to Gary Ka iser at P W R C for his support throughout. Funding was provided by the Nestucca Oil Spil l Trust Fund and Environment Canada . Permiss ion to work in Duu Guusd Tribal Park was kindly granted by the Old Massett Vi l lage Counc i l . Car l Ze i s s C a n a d a provided one pair of 10 x 40 binoculars for use throughout this project. I would like to acknowledge and thank those that were valuable co-workers in the field. Bruce Fitz-Earle shared his enthus iasm and humour, outdoor and cooking skil ls over three long field seasons . Brent Matsuda and Tara Burke are thanked for their enthus iasm, sense of humour, and hard work in 1995. My appreciat ion and s incere thanks for many nights of shrewing and days of eagl ing that appeared to never end. Mark Drever and Rowley Taylor ass isted with the radio-collaring of rats and shared their expertise in Norway rat ecology. I thank Peter Buck for introducing me to the rich history of the islands, support in the field, and acting as a travel agent between Massett and Langara Island. J ames Hageman ass isted in many aspects of field work. Karen T imm at N W R C worked through many t issue samples in preparation for analys is. Dr. Malco lm Mcad ie and Tom Packer necropsied ravens and conducted protein electrophoresis respectively. M.S. Bhatti and Dr. J . Thompson taught me the fine art of blood sampl ing birds. I would a lso like to thank the Schweers and Holyoak famil ies at the Langara Island Lightstation for opening their homes for us to dry off, shower, do laundry and eat a home cooked meal . Lynne Holland provided housing and a place for sorting, shipping, and c leaning of equipment in Massett. Most importantly, I am most grateful to my wife Kathryn, for her support and patience, and our son Douglas, for whom life is just beginning. xiv .to my parents for the many years of love and encouragement. x v Chapter 1. General Introduction The introduction of rats (Rattus spp.) to ocean ic is lands can have significant consequences for local wildlife populations, particularly for burrow-nesting seabirds which can decl ine in abundance or eventually be extirpated (Moors and Atk inson 1984). Introduced rats have been implicated as a significant factor in the dec rease or extirpation of the Langara Island, British Co lumbia, Canada (see sect ion 1.2 and Figure 1.2) breeding population of seabirds that was once descr ibed as " immense" and "astronomical" (Drent and Guiguet 1961). Seab i rds that previously utilised Langara as a breeding site included: tufted puffins (Lunda cirrhata), Leach 's storm petrels (Oceanodroma leucorha), fork-tailed storm petrels (O. furcata), Cass in ' s auklets (Ptychoramphus aleutica), rhinoceros auklets (Cerorhinca monocerata), and ancient murrelets (Synthliboramphus antiquus) (Campbel l et al. 1990). But, s ince the 1950's, the breeding populations on the Island have decl ined significantly or have been extirpated. Cass in ' s and rhinoceros auklets and Leach 's and fork-tailed storm petrels no longer breed on Langara, while only a small breeding population of tufted puffins exists on nearby Cox Island (Taylor and Ka iser 1993). In 1993, the nesting population of ancient murrelets were estimated to be less than 10% of historical numbers, and have decl ined by 4 0 % since the late 1980's (Harfenist 1993; Bertram 1989). The ship rat (Rattus rattus), likely introduced during the fur trade in the early 1800's, was present on Langara Island, but has been d isp laced by the Norway rat (Rattus norvegicus), likely introduced in the 1940's. The introduction of the Norway 1 rat co inc ided with the decl ine in the seabird colony (Taylor 1993), and there is ev idence for significant rat predation on murrelet eggs, chicks, and adults (Bertram and Nagorsen 1995, Harfenist 1993, Bertram 1989). Simi lar problems have been reported elsewhere, but s ince 1981, New Zea land biologists have successful ly eradicated introduced rats from smal l ocean i c is lands with the use of anticoagulant rodenticides. Recent rat eradication programs in New Zea land used the anticoagulant brodifacoum, d ispensed from fixed, evenly spaced bait stations on offshore is lands (Taylor and Thomas 1993; 1989). In 1995, the Canad ian Wildlife Serv ice attempted the complete eradication of Norway rats from Langara Island and adjacent Cox and Lucy Islands with the application of the second generation anticoagulant, brodifacoum ( IUPAC 3-[3-(4'-bromobiphenyl-4-yl)-1,2,3,4-tetrahydro-1-naphthyl]-4-hydroxycoumarin ) (Figure 1.1), using a technique developed in New Zea land (Taylor and Ka iser 1993). This was the first time this method (Section 1.3) was employed in North Amer i ca and it could have potential for other is lands in the Queen Charlotte archipelago (Bertram and Nagorsen 1995). However, the use of brodifacoum to eradicate rats on offshore is lands poses a risk of primary and secondary poisoning to non-target spec ies . Pr imary poisoning results when the bait and anticoagulant are consumed directly by a non-target animal. Secondary poisoning occurs when a primarily poisoned animal is consumed by a predator or scavenger (Colvin et al. 1988). 2 T3 C (D .TO GO E o o •2 2 I O ro c ro 13 O o z, £ ro o 8i> ro o .2 c E .ro ro o o E E o o E (D >« W D) C o d) ro M . o E £ 8 2 2 4 -Q (D O ±= (D (U ZJ O •4-1 o ^ ~ ZJ ro o o Q. E 5 (D CL 6 § to • ' i _ ^ ro T~ CL E 3 8 t. .P 3 1.1 Objectives of Research The objective of the Canad ian Wildlife Serv ice was to examine the feasibil ity of eradicating introduced rodents from seabird colonies by balancing the long term benefits against the relative costs. The benefits are the removal of a significant, exotic predator spec ies from islands which are important seabird breeding areas along the Pacif ic Coast . The costs, other than financial, include the effects on the native avian and mammal ian spec ies that live on and around these seabird colonies. There were three main areas of poisoning concern for non-target spec ies: 1) Langara Island has a native population of dusky shrews (Sorex monticolus elassodon Osgood) . Deer mice (Peromyscus maniculatus) were also known to once inhabit the is land. The New Zea land baiting model was des igned to ensure eradication of the introduced rats and mice because there were no native smal l mammals to be concerned about. 2) Rats are known to die above ground after ingesting a lethal dose of an anticoagulant (Cox 1990) posing a possible secondary poisoning hazard for native predators. However, few researchers have investigated the actual proportion of rats dying above ground, and the importance of rats in the diet of the local predators and scavengers on Langara Island was unknown. 3) Invertebrates are important in eliminating excess bait and poisoned ca r casses from the environment, however, they may introduce brodifacoum into the food chain, and pose a possib le secondary or tertiary poisoning risk. The overall goal of this thesis was to investigate the risk of non-target spec ies poisoning with the use of brodifacoum to eradicate rats from Langara Island. There 4 were three main objectives of my study: 1) monitor the native smal l mammal population over the course of the baiting. 2) evaluate the risk of secondary poisoning to avian scavengers and predators from poisoned rats, and 3) a s s e s s the invertebrates as a source of introduction of brodifacoum into the ecosys tem. In Chapter 2, I summar ise the mode of action of anticoagulants and the physiological basis for the non-target spec ies poisoning concern. In Chapter 3,1 examine the results of the endemic small mammal population monitoring before and after the intensive baiting period (Section 1.3). In Chapter 4, I examine the role played by rats in the secondary poisoning hazard to the local avian predators and scavengers . In Chapter 5, I determine the transfer of brodifacoum into the ecosystem, particularly by invertebrates consuming bait and poisoned rat ca rcasses . In Chapter 6,1 summar ise the results from the studies and I propose recommendat ions for minimising non-target primary and secondary poisoning hazards when eradicating rats from seabird colonies along the British Co lumbia coast. I a lso conducted two laboratory experiments to assist in the evaluation of data col lected in Chapter 4 and these are descr ibed in Append ix A. Environmental aspects of brodifacoum are presented in the appendices. 1.2 Study Areas 1.2.1 Langara Island Langara Island (54° 14'N, 133° W), a lso known as Kiis Gwai i or North Island, is located at the northwestern tip of the Queen Charlotte archipelago (Haida Gwai i), 5 British Co lumbia , Canada (Figure 1.2). The Island is 3300 ha in s ize and is relatively flat rising to a maximum elevation of 160 m. The shorel ine is highly variable ranging from rocky and sandy beaches to steep cliffs and bluffs. The Island is ringed by three bands of dominant vegetation that include: sitka spruce (Picea sitchensis) predominating along the shoreline, western hemlock (Tsuga heterophylla), and western red cedar (Thuja plicata) dominating the interior. Thick growths of salal (Gaultheria shallon) are abundant in areas, while Nootka reed grass (Calmagrostis nutkaensis) is found along the shorel ine. The forest floor is predominantly open, and covered with moss, and moss covered logs and stumps. Ra ised bogs are found inland amongst the lakes. There were two areas at which people reside year-round on the Island. At Langara Point, on the north-west tip, is the Langara Lightstation with two famil ies living permanently (Figure 1.3). At the southern most tip of Langara Island were the fishing lodges located in and around Henslung Cove . There were two permanent land-based lodges, West Coas t Fishing C lub and Langara Island Lodge on Iphigenia Point. Three other, seasona l , floating fishing lodges are barged in early spring and leave by early fall. Float planes regularly drop off and pick up tourists throughout the fishing season . 1.2.2 Lucy Island Lucy Island located in c lose proximity to Langara Island (about 300m) is 40 ha in area, 1400 m long and 300 m across at its widest. The shorel ine is highly variable consist ing of coarse gravel beaches and boulders to rocky she lves and outcrops. The island is relatively flat with a maximum elevation of 69 m above 6 7 8 sea level. Old growth forest predominates, but with heavy patches of windfall containing thick second growth running along the central portion of the Island. The dominant vegetation is similar to that found on Langara. 1.3 Baiting Protocol The following is a summary of the baiting protocol of the Langara Island Seab i rd Habitat Restoration Project. The protocol is descr ibed in detail in Ka i ser et al . (1997). The baiting was carried out with up to 70 people working out of five field camps d ispersed around Langara Island. The bait stations were deployed approximately every 100m and fastened down with two 60 cm long wires (Figure 1.4). The bait blocks measured approximately 3.5 cm x 3.5 cm x 2 c m , and weighed 20 g. The bait consisted of brodifacoum (Ratak+ ™) at a concentration of 0.005%, or 1 mg brodifacoum per bait block. The carriers and attractants were a mixture of blood, bone, wheat, tallow and castor sugar (Kaiser et al. 1997). Paraffin wax bound the bait together and "weatherproofed" the block. The baiting protocol was tested on Lucy Island beginning July 12, 1994 (Buck 1995). Ea ch station was armed with three bait blocks. The stations were checked each day and all activity was recorded. Activity included: bait d isappearance, chew marks by rats, shrews or invertebrates. Baits chewed by shrews or s lugs were destroyed and replaced. On day 19, baits were placed into plastic bags to minimise exposure to non-target spec ies. On August 17, 1994, all baits were removed from the stations. 9 E o H P 0 > o E 0 a: H g •+-< CD O TO CD « 2 ^ •pap , » i l p i i l » l M M TO c 0 CL o _ l M M mm CD o o •F E o o UO T3 0 Q. h= Q. O cz _ CO o " -o cu — TO a) i f o >+- CO O c CD O "co "o cz CO cu Q . CO cu sz ZZ CO .5" cu c o Z3 . c SZ CO -*—» 1— > o "D cu ca JS = .2 E a) o J t: o CU TO *:§ o cu CO cu co CO co c o 0 CO cu . i i SZ > co CO CO •4—* CO D) _ tz CU CO SZ o co o co LU E o CO "O O ila_ X CU CO o o O co . o CO CD 0 , 5 CN 0 "O iS o TO 0 co £ V c CO 0 CO -Q TO CO c > CO o - 1 E > ; CD = i o TD T3 ~ 0 0 3£ c TO 0 CO 0 c oc -t—' CO o E T3 c CO L_ 0 SZ "co 0 3R 0 0 CO cz o "co -4-* co Q . Q. 0 TO CO cz 'S CO 0 CO 2> 0 $ CO c o » CO CO 0 SZ o E o M— T3 0 CL co T3 CO i — _ 0 3^ o .o c 0 -CO • TO • i — £ 8 = o .s>> Li. Q _ 0 _ E ® 2 | ^ 2 § cu 0)'+; c ° n ° Q.-i= 0 n cz a id 0 co C -D Q_ 1 0 Rats were d iscovered again on Lucy Island in late August 1994, and in 1995, the Island was re-baited in attempt to eradicate the remaining rats. On July 11 1995, intensive baiting began on Langara Island. Each of the 3848 bait stations (Figure 1.5) were armed with 6 to12 bait blocks , monitored every 2 to 3 days, and replenished as necessary until early August. For the post-intensive baiting period between August 1995 and August 1997, each station on Langara Lucy and C o x Islands was armed with three bait blocks wrapped in plastic produce bags and placed on an aluminium or plastic tray to keep the bait dry. The bait stations were checked for ev idence of rat use in September 1995, and again in February, May, and August 1996. The stations and bait were removed from the Islands during the summer of 1997. 11 LANGARA ISLAND Figure 1.5. Bait station Layout on Langara, Lucy and Cox Islands, 1995. Each dot represents 1 bait station. Total number of bait stations: 3848. - J&i L* - • • i cox is. "£Jb (:i^ ^u:\::u 4^4v^ .7h .Jj£l • •"• t: ?: •::)) *m sit - * • — * _* • _ • « i « J si: *.::::*• ":\ •\ v . • ' J 0.5 Kilometers LUCY IS. 12 Chapter 2. Hemostasis and Mode of Action of the Anticoagulant Rodenticides: the Physiological Basis for Non-Target Species Poisoning Concern. 2.1 Introduction Hemostas is is one of the many physiological characterist ics that is shared by birds and mammals . Hemostas is has two major functions: (1) to prevent blood loss from sites of vascu lar disruption and (2) prevent pathologic thrombosis (excess clotting) by limiting clot formation to sites of vascular disruption (Brandt 1991). Hemostas is can be v iewed as two forces which continuously oppose and ba lance each other. Any shift in the balance of forces will result in excess thrombosis or bleeding. 2.2 Hemostasis Under balanced conditions, hemorrhaging is controlled by clot formation which results from a ser ies of proenzyme to enzyme transformations culminating in the convers ion of fibrinogen to fibrin (Kase et al. 1980). In mammals and birds, this coagulat ion or enzyme cascade (Macfarlane 1964) can be activated by one of two ways: extrinsically or intrinsically. Intrinsic activation involves contact activation such as the exposure of the blood to the subendothel ium in vivo or, contact of blood to a negatively charged surface such as g lass in vitro (Macfarlane 1964; Rapaport 1987). Recently, the well studied human coagulation system has revealed that contact activation is not required for normal hemostas is (Jesty and Nemerson 1995). Past studies have indicated that birds lack an effective intrinsic clotting mechan ism (Didisheim et a l . 1959). However, Doerr and Hamilton (1981) have provided ev idence of a 13 functioning intrinsic mechan ism in chickens, but this plays a minor role in the overal l coagulat ion function in compar ison to the extrinsic pathway. This is similar to the accepted human model in that the coagulation cascade is initiated by t issue factor (thromboplastin) and not necessar i ly the intrinsic pathway (Jesty and Nemerson 1995). The extrinsic pathway involves a t issue factor (or thromboplastin) that normally res ides in the endothel ium and other t issues such as the brain (Griminger 1986). The thromboplastin is present on the surface of many cell types and is not normally in contact with the circulating blood (Jesty and Nemerson 1995). Upon t issue damage, the thromboplastin, now exposed to blood, binds with factor VII forming an enzymatical ly active complex which initiates the clotting cascade (Rapaport, 1987) (Figure 2.1). The mammal ian and avian extrinsic pathway coagulation cas cades are functionally similar (Griminger 1986; Bellevil le et al. 1982; Ka se 1978). Once the factor Vl l-thromboplast in complex forms, it initiates the activation of factor X which provides a positive feedback increasing the enzymat ic activity of the factor V l l -thromboplastin complex (Rapaport 1987). This complex along with activated factor XI then activates factor IX. The activated factor Vl l-thromboplast in and factor IX complexes are required to activate sufficient factor X to generate enough thrombin to maintain hemostas is (Rapaport 1987). Thrombin (activated factor II) transforms fibrinogen into fibrin followed by the stabilization of the fibrin clot. Interruption of this cascade at any step will prevent the formation of the fibrin clot and hemorrhaging may continue uncontrol led. The introduction of an anticoagulant, such as 14 XI • Xla Intrinsic Pathway Extrinsic Pathway IX Ca++ Vila IXa Ca++ Ca++ Xa Ca++ Prothrombin Thrombin Thromboplastin Ca++ VII XIII Xllla IA Ca++ Fibrinogen Fibrin ' 1B Stabilised Fibrin Figure 2.1. A diagrammatic representation of the coagulation cascade . The grey arrows refer to the intrinsic pathway. The roman numerals represent the clotting factors. The lower case a, represents the activated form of that clotting factor. Dashed lines represent positive feedback loop enhanc ing the coagulat ion cascade (adapted from Sturkie 1986 and Rapaport 1987). 15 brodifacoum, in sufficient doses, will result in the inhibition of production of some clotting factors required in the extrinsic pathway. 2.3 Vitamin K 1 and the Mode of Action of the Anticoagulants The metabol ic role of vitamin K., is to activate the vitamin K dependent clotting factors- II (prothrombin), VII, IX, and X (Rapaport, 1987; Suttie 1980) by contributing a carboxy side chain by post-translational carboxylation of se lected glutamic res idues to form gamma-carboxyglutamic acid residues (Figure 2.2). These amino acid residues are necessary to chelate divalent ca lc ium to interact with phosphol ipid containing membranes, their normal site of activation (Goodman et al., 1985). Vitamin K-, now in its oxide form, can be reduced to its original form by the epoxide reductase enzyme. The recycled vitamin K: is now avai lable for further activation of the vitamin K dependent clotting factors. The anticoagulants interfere with this process, resulting in the release of inactive clotting factors into the blood stream. Al l coumarins, the group of anticoagulants to which brodifacoum belongs, have the same mode of action. Ant icoagulants bind to the warfarin binding protein (Thijssen and Baars 1989) in the endoplasmic reticulum of the liver hepatocytes (Searcey et a l . 1977) and inhibit the epoxide reductase enzyme preventing the reduction of vitamin K 1 -epoxide. The 4-hydroxycoumarins do not necessar i ly bind to the reductase enzyme, but it has been speculated that they interact with a subunit structure with which the enzyme has to interact for normal functioning (Thijssen and Baars 1989). The inihibition of the reductase enzyme leads to a build up of hepatic 16 T3 0 0 sz .ti .Q 0 .E <o 0 CO CO o ZJ "O 0 c -c -~ co •— 0 "co o c CO oCZ 0 E >» N CZ 0 0 CO o co id CO C O * •* o O Q. _C0 0 e d> « l I I CO _Q O J= o 8 E co CO c TO O) co § .52 S O s °-co 0 2 CO >, 0 -o X 0 o .Q CD I* •2 E x : co 0 E o co 0 i !c cz o "co w X 0 co _£Z •4—• CO CL 0 > co tz Q. CO 0 i— CO ^ 0 S£ 'I -Hi I o co co 0 -E •i ° E ~ a s > 'sz 0 c c csi -I CN o) a% CO - o N CZ 0 0 CO co o ZJ "O 0 T3 CD ZJ sn T3 CZ co CO 0 o -I—» CO cz o _ co CO CD 3 o 0 •g X 8. E cu co 0 -•~ CO O - cz DO q Q- 0 £ Q. co o CO "O .E Q. CO i -E o ZJ O CO O T3 ^ CO 0 CZ co 0 | CN ? 00 O CD 0 Z3 CO CO -O w £ "co cz ^ ro 0 cn f= o o % 5 "cz c co § n T3 0 0 f < 0 "O c CD t 0 >p O N E 0 CL 17 vitamin K 1 epoxide (Leek and Park 1981; Caldwel l et al. 1974) and the level of activated clotting factor production decl ines or is inhibited (Choonara et a l . 1988; Leek and Park 1981). The extent to which the anticoagulant inhibits the production of the clotting factors is dose dependent (Thijssen and Baars 1989; 1987). In liver, there are specif ic, high affinity, saturable binding sites for anticoagulants (Thijssen and Baars 1989; Huckle e t a l . 1989a; Huckle et al. 1989b; Parmar et a l . 1987). The anticoagulants need to saturate the binding sites to initiate the anticoagulant effect (Parmar et al. 1987). Without the carboxylation of the vitamin K dependent clotting factors, they are re leased into the blood with little or no enzymat ic viability and the level of active factors decl ines. The anticoagulant effect develops gradually over time as the level of active clotting factors decl ines according to their different rates of biologic decay (Rapaport 1987). There exists a threshold of active clotting factors below which significant bleeding will result (Hoffman et al. 1988). Without active clotting factors, any trauma induced or spontaneous bleeding is uncontrollable and death results from hypoxia, and hypovolemic shock. All birds and mammals share the coagulation characterist ics that make them suscept ib le to anticoagulant rodenticides. The differences in sensitivity to anticoagulants results primarily from varying ability to metabol ise or excrete these compounds (Huckle et al. 1989b). In general, birds are less suscept ible to ant icoagulants as they are more readily able to metabol ise the compound while mammals are unable to metabol ise the anticoagulants before the lethal 18 anticoagulant effect takes place (Huckle et al. 1989b). Further d iscuss ion into potential sub-lethal and long term effects of brodifacoum exposure is presented Append ix C. 19 Chapter 3.0 The Short Term Impacts of Brodifacoum Baiting on the Native Small Mammals. 3.1 Introduction In North Amer ica , native small mammals are abundant and ubiquitous in a variety of habitats including offshore islands. The use of anticoagulants to eradicate introduced rats from these islands may alter the abundance and composit ion of the native smal l mammal populations. There is little information on how controll ing or eradicating rats affect resident native small mammal populations, particularly on offshore is lands. Historical records indicate that the dusky shrew and the deer mouse were the only native smal l mammals known to inhabit Langara Island (Foster 1965, I. Mctaggart Cowan , pers. comm.). Based on morphological measurements and pelage colour, the shrews on Langara Island have been included with the north Moresby Island race (elassodon) (Foster 1965). Deer mice were known to be present in the 1940's but had apparently been extirpated by the 1960's when Foster (1965) recorded shrews but no deer mice. No ev idence of deer mice was found during recent trapping campaigns (Harfenist 1993; Bertram 1989). The d isappearance of the deer mouse coincides with the introduction of the Norway rat (Taylor 1993). However, in September 1994, one deer mouse was trapped in Lord Bight on the west coast of Langara Island (C. French, pers. comm.) which may have represented a previously undetected population. The native smal l mammals are at risk of poisoning because they share many characterist ics with the target rodent spec ies. The risk of primary poisoning is 20 related to the palatability and availability of the bait in space and time. The bait consist ing of both animal and grain products may be attractive to both dusky shrews and deer mice. Dusky shrews are normally insectivorous, but are known to consume carrion (Cox 1990; V. Craig, UBC , pers. comm.). They are inquisitive and readily eat animal carcasses , showing a preference for organs, particularly the liver, in which anticoagulants accumulate (Cox 1990). The objective of this study was to identify the native smal l mammal spec ies at risk of poisoning, and monitor the short term population changes, abundance and composit ion, over the intensive baiting period. 21 3.2 Materials and Methods Non-target Small Mammal Identification and Population Monitoring In 1994, a 1ha grid of 49 trap stations (7x7), 15.2 m apart, with one Longworth-style live trap located within 2m of each station was establ ished on Lucy Island (Figure 3.1). Only one grid was used due to the unavailabil ity of traps, and no control grid was used. A three evening arming and check period was employed. The traps were baited with a mixture of peanut butter and oats, and coarse brown cotton was suppl ied for bedding. The traps were armed in the early evening and checked 5-6 hours later, at dusk or just after. This was repeated for the following two evenings. Live trapping was conducted once before and once after the baiting, in July 1994, 3 days before the start of the baiting and in August after the bait was removed from the stations. In May 1995 and August 1995, a single trap sess ion oo-the establ ished grid was performed. In 1995, Langara Island was divided into three treatment regions: the existing Anc ient Murrelet colony at Mcpherson Point (a high population of rats), the east coast in Eger ia Bay (lower rat population), and the west coast region in Lord Bight (possible remnant population of deer mice) (Figure 3.1). In each of the regions, two 1-ha grids, each with 49 trap stations (7x7), 15.2 m apart, were establ ished. Longworth style live traps were placed on all grids except for two which had smal l Shermann traps. Each grid was trapped twice before the baiting began and twice after the intensive baiting period. 22 23 Graham Island, 1 km across Parry Pa s sage from Langara Island, was used as a control site. Two control grids were establ ished, however, one was removed after cont inuous disruption by a female black bear and her two cubs during the second trap sess ion in June 1995. A three evening arming and check period was employed on Langara and G raham Islands following the same procedures used on Lucy Island. This short arming time was utilised because shrews die if left in traps over night (Sull ivan 1990). Trapping was performed in the evening to increase the chances of trapping any deer mice present on Langara or Lucy Islands. Each shrew captured was weighed with an Avinet spring balance, and marked for future identification with a blonde hair dye in a unique combination of dots on its fur and/or toe-nail c l ipped. I chose not to toe-clip because if any bait was consumed, lethal hemorrhaging may have been induced. Due to the difficulty in sexing juveni le or non-sexually active shrews, only obvious reproductive condition was noted (Craig 1995; Hawes 1977). Individuals were re leased immediately after process ing. The traps were locked open between trap sess ions, and also between checks and arming t imes. Shrews caught five or more t imes (Craig 1995) were used for estimation of range s ize and distance traveled calculated with the computer program C A L H O M E (Kie et a l . 1994). Range s ize was calculated using 90% minimum convex polygons (MCP ) for compar ison to other studies (Craig 1995; Hawes 1977). 24 Shrew use of Bait Stations Bait station operators were asked to record shrew activity in bait stations. Training in the identification of shrew chewed bait blocks was provided prior to the baiting. Shrew chews were identified from the s ize and pattern of the incisor marks. Any shrew chewed bait block was replaced to prevent it being recounted during future visits. However, only data col lected by identified, reliable bait station operators were used. Calculations and Statistical Analyses Population Size Changes The data are reported as the number of individuals/ha and unique shrews /100 trap nights (TN). Su c ce s s per 100 TN was calculated following Nelson and Clark (1973) as: CE= A x 100/ (TU - IS/2) Where CE= catch effort; A= number of animals; TU= trap units (TU= PxIxN where P= number of trapping intervals (3 nights); l= length of trapping intervals (1 night); and N=number of traps (49 traps); S= number of traps c losed and empty or with a recaptured shrew. Where possible, the Schnabe l /Schumacher-Eschmeyer method was used to estimate population s ize (Krebs 1989). Conf idence intervals were calculated using the Po isson distribution. The trap success data were square root transformed in an attempt to normal ize the data. I used a two way A N O V A for repeated measures (Kuehl 1994). 25 Ana lys is was carried out with the JMP statistical package ( S A S 1995) with the following statistical model: Yi J k = H +/?, + Tj + (R7V E«k where Yj j k = unique shrews/1 OOTN, Rt = effect of region, and Tj = effect of the jth time, and (f?7~)jj = the two way interaction between time and region, and E i j k = random error. A signif icance level of P<0.05 was chosen a priori. 26 3.3 Results Only dusky shrews (63 unique individuals) were trapped on Lucy Island in 1994 and 1995, and on Langara Island (182 unique individuals) in 1995. On G raham Island, deer mice (32 unique individuals) were abundant on control grids (Table 3.1), and dusky shrews (18 unique individuals) were also captured. Table 3.1. Abundance of deer mice (Peromyscus maniculatus) /100 Trap Nights on Graham Island, before and after the intensive baiting period on Langara Island, 1995. Before Baiting Post Baiting Location May July August Mid-August East Grid 7.9 Wes t Grid 8.8 21.0 17.4 4.3 Lucy Island In 1994, there was a decl ine in the number of unique shrews captured after the baiting (Figure 3.2, Tab le 3.2). By May 1995, trap success rebounded to half the pre-baiting 1994 estimate. Unique number of shrews captured again decl ined over the baiting period in 1995 (Table 3.2). The proportion of shrews in breeding condition decl ined in both years over the course of the baiting period (Table 3.3). 27 F i gu re 3.2. The number of unique dusky shrews (Sorex monticolus) captured on Lucy Island, 1994 and 1995. The vertical arrows indicate when the brodifacoum baiting for introduced rats began. The bait was removed in mid August, 1994, but was reappl ied in July 1995. Tab l e 3.2. Dusky shrew (Sorex monticolus) trap success (# shrews caught/100 Trap Nights) and population estimate before and after brodifacoum bait application Lucy Island, 1994/1995. Before Baiting Post Bait ing 1994 Trap Suc ce s s 19.9 2.8 Populat ion Est imate 56.4 N D a (32-155) b 1995 Trap Suc ce s s 11.8 10.5 Populat ion Est imate 28.1 13.2 (19-185) b (11-47) b a Not Determined. All shrews were trapped on the third evening, b 95% Confidence Interval 28 Table 3.3. Proportion of dusky shrews (Sorex monticolus) in breeding condition (testes scrotal, nipples large) before and after baiting on Lucy Island, 1994/1995 (n in brackets). Pre-Bait ing Post-Bait ing 1994 0.40 (10) 0.25(1) 1995 0.56 (10) 0.13(2) Langara Island There was no significant interaction between time and regions (Table 3.4). Overal l , there was no statistical difference in trap success before and after the intensive baiting period on Langara Island (Figure 3.3). Trap success was very low on the smal l Shermann trap grids (Table 3.4), but removing those grids from the analys is had no effect. May July August Mid-August Figure 3.3 Mean unique number of dusky shrews (Sorex monticolus) captured on Langara Island and on the control grids on G raham Island, 1995. Vert ical arrow indicates when brodifacoum baiting for introduced rats began on Langara Island. 29 Table 3.4. Dusky shrew (Sorex monticolus) trap success (# caught /100 Trap Nights) by trap sess ion and region on Langara Island and G raham Island, 1995. Before Baiting Post Bait ing Reg ion Grid Location May July August Mid-August Langara Island Murrelet Co lony No-Name Pt 14.3 1.4 a 6.9 17.1 Explorer Bay 3.4 9.1 6.8 3.6 East Coas t Eger ia North 3.0 11.1 2.8 7.0 Eger ia South 3.5 3.5 1.4 0.7 Wes t Coas t Lord Bight N o r t h a 0.7 1.4 0.0 1.4 Lord Bight South 12.5 10.3 3.5 3.5 G r aham Island East 2.4 _ _ _ (Control) Wes t 0.0 5.7 5.7 1.1 a. Small Shermann traps used on these grids. Whi le statistically not significant, the regional number of unique shrews captured per hectare decl ined between the start of the baiting and the third trap sess ion in all regions on Langara Island. The greatest decl ine occurred on the east coast, dropping by 7 5 % (Figure 3.4). On the G raham Island control grid, there was an accumulated trap mortality of three shrews during the second and third trap sess ion. During the last trap sess ion, 31, 35 and 36 of the 49 traps over the three trap nights, on the grid were reported c losed and empty. 30 Ancient Murrelet Colony May July August Mid-August East Coas t May July August Mid-August Wes t Coast May July August Mid-August Figure 3.4. Unique number of dusky shrews {Sorex monticolus) captured on each study grid in the three regions on Langara Island, 1995. Vertical arrows indicate when brodifacoum baiting began. Longworth traps were used on all grids except where denoted by a * when small Shermann traps were used. 31 The proportion of shrews captured in breeding condition did not change on the control grid and remained at 50% (Table 3.5). However, the proportion of shrews in breeding condition decl ined at all three baited regions, with the most drastic decl ine (%) occurring on the Wes t Coas t (Table 3.5). . Table 3.5. Proportion of dusky shrews {Sorex monticolus) in breeding condit ion before and after intensive baiting, Langara Island, 1995 (mean of two grids per region; sample s ize in brackets). Murrelet East Coas t Wes t Coas t G raham Co lony Island (Control) Pre-Bait ing 0.44 (18) 0.43 (7) 0 .49(18) 0.50 (5) Post-Bait ing 0.39 (17) 0.21 (4) 0.06 (1) 0.50 (4) Shrew Use of Bait Stations Shrew incisor marks on bait blocks were smal l , and readily identifiable. The shrews chewed the edges of the blocks, working their way around all edges, never consuming the whole block. Conversely, rat incisor marks were larger and they would begin and continue to chew on one side of the bait block. On Lucy Island shrews chewed 38 bait blocks in 2 8 % of the bait stations in 1994 (Buck 1995). Shrews were also photographed with automatic cameras exiting bait stations. By day 20 of the baiting program on Lucy Island in 1995, shrews had visited 8 0 % of 42 bait stations and chewed one or more bait blocks in each station. In Hens lung Cove , shrews had chewed 157 bait blocks in 50% of the 28 stations in that region over the course of the intensive baiting period. In the interior of the 32 Island, shrews had chewed 151 bait blocks in 80% of 53 stations visited. On Langara Island, bait station operators reported shrews "stumbling" out of stations, indicating possible anticoagulant intoxication (Cox 1990). A s well, during checks of the live traps on the North Eger ia grid in early August, smal l , blue colored shrew feces were noted in one trap indicating that at least one shrew had been feeding on the bait. Shrew Ranging Distances Only four individual shrews were captured five or more t imes, with the number of recaptures ranging from five to eight. The shrew 9 0 % M C P range s ize varied between 112 m 2 to 2700 m 2 . The maximum distances traveled between traps ranged from 61.8 m to 91.2 m (Table 3.6). Table 3.6. Number of captures, weight, maximum distance traveled and 9 0 % M C P range s ize for dusky shrews (Sorex monticolus) caught five or more t imes, Langara Island, 1995. Sh rew Locat ion Number Weight Mean Max imum 9 0 % Range No. of g Distance Distance A rea Captures m m m 2 Murrelet Co lony 8 8.3 40.3 90.2 2700.0 5 Murrelet Co lony 6 6.8 54.9 91.2 1238.0 6 Lord Bight 5 7.8 75.7 76.5 2475.0 32 Murrelet Co lony 5 5.8 54.6 61.8 112.5 33 3.4 Discussion Shrews were abundant in the coastal region of Langara Island prior to the baiting, but there was no evidence of deer mice. The results suggest that the dusky shrew was able to co-exist with introduced rats, perhaps because of niche dif ferences resulting in low interspecific competit ion. Rats may have avoided preying on shrews because they secrete a distinctive pungent, acrid odor making them unpalatable (Churchfield 1990; Hawes 1976). Deer mice were apparently extirpated from the coastal region of Langara Island either through predation, competit ion or both. Other is lands in the Queen Charlotte archipelago contain populat ions of deer mice except where there are rats; the exception are the larger is lands, Graham, Moresby, Kunghit and Lyell Island that have populations of both deer mice and rats (Foster 1965). The individual deer mouse trapped in Lord Bight in 1994 may have represented a remnant population, or may have been accidental ly transported from Graham Island by the survey crew. Despite use of bait stations by shrews, and decl ines over the course of the baiting, there were no significant short-term impacts on shrew abundance on Langara Island. Rat control programs have been shown to significantly impact other non-target smal l mammal spec ies. During anticoagulant poisoning for rat control in the spring and summer on farms in England, non-target woodmice (Apodemus sylvaticus) entered stations and fed on bait (Cox and Smith 1990). There was a 7 5 % reduction in the population s ize of woodmice after treatment while the control populat ions increased (Cox and Smith 1990). The individual survivorship of marked 34 individuals was between 0-19% on baited farms and 50% on control farms (Cox and Smith 1990). No est imates of survivorship could be made for the shrews on Langara Island. Accord ing to the trapping record, the G raham Island (control grid) dusky shrew abundance decreased while the Langara abundance increased between the third and fourth trap sess ion. It was expected that the reverse would be measured, i.e., shrews on Langara Island would decrease in abundance as compared to the control site. However, on the control grid, the trapping of non-target deer mice, high number of c losed but empty traps and trap mortality of shrews could account for the low number of shrews trapped on this grid during the third and fourth trap sess ion . The Lucy Island population of shrews decl ined sharply over the course of the baiting in 1994 and was the reason a monitoring program was establ ished on Langara Island in 1995. In 1995, shrews on Lucy Island did not decl ine in abundance as in 1994, but the proportion of shrews in breeding condition fol lowed a similar pattern of decl ine. On Langara Island, the greater the regional decl ine in number of shrews captured after the intensive baiting, the greater the decl ine in the proportion of shrews in breeding condition (Figure 3.5). There was no change in the proportion of breeders on the control grid over the same time period. This decl ine in the proportion of shrews in breeding condition may be due to seasona l changes such as recruitment of juveni les or it may represent a specif ic poisoning impact on shrews in breeding condition. This seems to suggest that the individual risk of primary poisoning was greater for individuals in breeding condition. A s these 35 0.5 0.45 0.4 1 0.35 0.3 Graham Island (Control) Murrelet Colony # •- 0.25 S 0.2 Lucy Island 1994 East Coast 0.15 0.1 I 0.05 West Coast • Lucy Island 1995 0.2 0.4 0.6 0.8 1 Post Baiting Shrew Population as a Proportion of Pre-Baiting Abundance Figure 3.5. The proportion of dusky shrews (Sorex monticolus) in breeding condition correlated to the post baiting shrew population estimate as a proportion of the pre-baiting population estimate (r = 0.64). territories were vacated, dispersing juveni les and other non-breeding shrews may have competed to fill those vacant territories (Churchfield 1990). This would lead to an increasing density of shrews which may explain the increased population during the fourth trap sess ion in mid-August on four of the six grids on Langara Island. The risk of primary poisoning to shrews from gaining access to the bait in stations is greater for shrews in breeding condition because they range more than non-breeders (Hawes 1977). The greater ranging distances of breeding shrews increased the chance of encountering brodifacoum in stations. Breeding males may 36 be at greater primary poisoning risk than the breeding females, non-breeders and juveni les, as they range significantly further afield (Hawes 1977). A n extreme movement of one scrotal adult male was noted in Lord Bight. This individual was live-trapped in a field camp, marked and released, and was subsequent ly recaptured 7 hours later on the North Lord Bight grid, 181.5 m away- almost twice the average distance between bait stations. The range s izes calculated for dusky shrews on Langara Island were smaller than the ranges calculated by Hawes (1977) but similar to those found by Craig (1995). The distance between stations was adequate to spatially exclude some shrews. This study addressed the short term impacts of brodifacoum baiting on the shrew population. What remains unknown is the impact of having bait left in bait stations wrapped in plastic bags for the 2 year period after this study ended. It can be conc luded that shrews were attracted to bait in the stations and that there were short term impacts on shrews including breeding individuals. However, they were not extirpated from Langara or Lucy Island. Shrew trap success rebounded 9 months after bait was removed from Lucy Island. Thus, cessat ion of the baiting could al low for the shrew population to increase to pre-baiting levels within 1 year. The long term impacts of brodifacoum baiting on the non-target dusky shrew would require further investigation; however, there are unlikely to be any significant long term impacts. 37 Chapter 4. An Evaluation of the Secondary Poisoning Hazard to Avian Wildlife. 4.1 Introduction The secondary poisoning risk to birds from feeding on anticoagulant killed rodents is well known and has been demonstrated in the laboratory (Newton et a l . 1990; Radvany i et al. 1988; Townsend et al. 1981; Mendenhal l and Pank 1980). A s ses s i ng the actual hazard under field conditions is difficult because pharmacologica l susceptibil ity is not necessari ly an indicator of ecological susceptibil ity (Moore 1966) and predators and scavengers are not expected to consume only contaminated animals (Townsend et al. 1984). The risk of secondary poisoning to avian predators and scavengers from brodifacoum poisoned Norway rats is related to exposure factors such as the behaviour of the target spec ies during the latent period, the location of ca rcasses (above or below ground), the anticoagulant residue loading in the target spec ies, and the behaviour and diet of the non-target spec ies (Record and Marsh 1988; Taylor 1993; Kauke inen 1982). Ant icoagulant poisoned Norway rats demonstrate altered behaviour which potentially makes them more susceptible to predation and scavenging. Norway rats exposed to a lethal dose of an unidentified anticoagulant showed reduced thigmotactic behaviour (moving in contact with a vertical surface such as a wall) and spent significantly more time in open areas than under cover as compared to the controls (Cox and Smith 1992). Whi le in the open, rats were observed sitting motionless or staggering about shortly before death. On death, 50% died in open 38 areas, apparently deliberately moving out of nest boxes into open areas just before dying. During the field testing of the baiting protocol on Lucy Island in 1994, one of three brodifacoum poisoned radio-collared Norway rats died above ground away from a burrow (Howald 1995). Ca r cas ses of poisoned rats also were found on the beach and other open areas under the forest canopy. Secondary poisoning of non-target spec ies from the use of brodifacoum to control rats and voles has been demonstrated in the field. Eastern screech-owls (Otus asio) died after exposure to brodifacoum after broadcast application to control vo les in orchards (Colvin and Hegdal 1988; Merson et al. 1984). In New Zea land , Western weka (Gallirallus australis australis) and Stewart Island weka (Gallirallus australis scotti) died after consuming rats that had fed on bait containing brodifacoum (Taylor 1984). C ommon ravens (Corvus corax), Northwestern crows (Corvus caurinus) and bald eag les (Haliaeetus leucocephalus) were considered to be at risk of secondary poisoning from feeding on dead or dying Norway rats on Langara Island (Taylor and Ka iser 1993). The importance of rats in their diets is unknown, but, they are opportunistic scavengers and predators that may take advantage of a new prey source. To evaluate the secondary poisoning risk of the three avian spec ies, an integrated study was undertaken to determine the extent to which rats will die above ground, and to evaluate the poisoning of non-target spec ies. The main objective 39 was to evaluate the secondary poisoning risk to avian predators and scavengers from the use of brodifacoum to control Norway rats on Langara Island. The specif ic hypothesis tested was that dying or dead rats would be avai lable to predators and scavengers , thus putting them at risk of secondary poisoning. 40 4.2 Materials and Methods 4.2.1 Carcass Locations of Brodifacoum Poisoned Norway Rats Fifty Tomahawk live traps were armed for 400 trap nights & 400 trap d in early Ju ly 1995, between Mcpherson Point and No-Name Point, in an attempt to capture 5 rats of each age (juvenile/adult) and sex c lass. Trapped rats were anaesthet ised with halothane, sexed, weighed with a spring balance, ear tagged with a fingerling tag, and fitted with a PD-2C radio collar (Holohil Sys tems, Ontario, Canada) . Four adult males, five adult females, five juvenile males and five juveni le females were radio-collared between three d before the start of the program to the day the intensive baiting began. One adult male and one adult female were trapped on day 2 of the program. Each rat was located at least once per day by tracking the signal to its location and taking a bearing and measurement to the nearest bait station. At any one time, a rat was determined to be alive if, when holding the antenna steady, the signal strength became weaker or stronger indicating that the animal was moving, or if the rat was found to be alive at a future time point. The time to death was calculated as the number of d e lapsed between the start of the baiting program and the date of last known activity +24 h. 4.2.2 Brodifacoum Residues in Norway Rats Found Dead Above Ground Rats found dead by bait station operators and research staff were sexed , weighed and frozen in pre-labelled whirl-pak bags. Five adult male, four adult 41 female, and three juvenile male rats were sent to the National Wildlife Resea rch Centre, Hull, P Q for t issue preparation for brodifacoum residue analys is. The liver, GIT (gastrointestinal tract, including its contents), and the carcass were individually homogen ised and analysed for brodifacoum as descr ibed below. The data were log transformed (X|= log 1 0(x+1)). Ana lys is of the data was carried out using the JMP statistical package (SAS , 1995) with a two way A N O V A . The statistical model was: Yj J k = u. + A + Tj + (AT)ij + E i j k where Y i J k = log brodifacoum residue measured, A, = the effect of the ith age/sex c lass, and 7j = effect of the jth t issue, (AT)jj = the two-way interaction between dose effect and time of blood collection, and E i j k = random error. 4.2.3 Norway Rat Scavenger Identification In 1994, a year before the baiting operation began, non-poisoned, snap-trapped rats were laid out in open, exposed areas such as on beaches, open areas under the forest canopy, and around bald eagle nesting trees. Automatic, infra-red motion sens ing cameras were used to identify scavengers . A total of 29 rats were put out, 18 with cameras, and 11 on Eger ia Bay beach where scaveng ing spec ies were identified from tracks left behind on moist sand, or by direct observat ion. T ime to scaveng ing was only roughly estimated between visits to carcass sites or from direct observat ion. 42 4.2.4 Effects on Predators and Scavengers 4.2.4.1 Common Ravens All bait station operators, research and Langara Lightstation staff were briefed before the baiting program began and were encouraged to report observat ions involving ravens, or to turn in any carcasses , pellets, or other remains they found. Such ca rcasses were labelled and frozen for necropsy at the University of British Co lumbia , Department of An imal Sc ience by Dr. Ma lco lm Mcadie, a veterinarian exper ienced in examining pesticide poisoned birds. Livers were removed and frozen for brodifacoum residue analysis. Source of Brodifacoum Poisoning to Ravens The source of brodifacoum was determined from protein electrophoresis of the g izzard contents and by evaluation of the gizzard and intestinal contents for Norway rat hairs and bait fragments. Each gizzard was opened along its length and the contents, if any, were removed and the cutica gastr ica rinsed with water from a squeeze bottle. If there was an adequate amount, a sample was frozen and sent with control samples of Norway rat, shrew, raven, snails, and bait to the Alberta Natural Resources Service, Enforcement-Field Serv ices, Forens ic Laboratory, Edmonton, Alberta for identification by polyacrylamide gel electrophoresis (McClymont et al. 1982). The remaining gizzard contents were rinsed through two layers of cheesec loth and a 60 mesh s ieve lined with filter paper before using a dissect ing microscope (7-30x) to identify remains and estimated percent per vo lume. 43 The intestines were cut into three sect ions and the contents squeezed into a pre-washed g lass dish. The contents were examined for hair or bait crumbs under a dissect ing microscope. All hair was identified to spec ies under a microscope by compar ison against control samples of Norway rat, dusky shrew, black-tailed deer (Odocoileus hemionus columbianus), human and publ ished micrographs of deer, shrew, and Norway rat (Moore et al. 1974; Adorjan 1969). Raven Activity-1996 In Apri l 1996, the potential ongoing raven exposure to brodifacoum on Langara Island was investigated. Formerly active nest sites were visited at Mcpherson Point and Hazardous Cove . New nests were located by walking up to 200 m inland parallel to the shoreline; one observer between the beach and cliff bottom, the other along the ridge that rings the island. Regurgitated pellets and prey remains were col lected from under and around the nest sites. A total of 107 bait stations reported active (baits removed and/or stations disrupted) were visited. The plastic bags, a luminum trays used to keep the bait off the floor of the station, and bait remains were collected and examined for beak marks. Plast ic bags were examined under a dissecting microscope and compared against reference examples of raven-torn plastic bags. Bags were examined for stress bars and/or beak marks. 44 4.2.4.2 Northwestern Crows Crows (25) were collected by shotgun between May and August 1995. Each crow was weighed and sexed, and selected morphological measurements were taken before the livers were removed and placed into pre-labelled whirl-pak bags and frozen. The livers were pooled for brodifacoum analysis based on collection dates and/or location. Control livers from pre-baiting snap-trapped rats were a lso sent for brodifacoum analysis. 4.2.4.3 Bald Eagles Using a fish snare (Jackman et al. 1993; Ca in and Hodges 1989), bald eag les were trapped and a blood sample drawn between days 7 and 47 of the intensive baiting campaign (Table 4.1 and Table 4.2). Three eagles were trapped prior to the start of the eradication in order to test the trapping methods and to obtain control bipod. Trapping was concentrated where eagles commonly roosted on the coast, and ranged from the Langara Island Lightstation east to Mcpherson Point, south along the east coast to Holland Point, and west to Cox Island. Trapping was attempted, but unsuccessfu l , on the west coast due to the incessant northwest/westerly ocean swell which hindered trapping. W e commonly observed eag les flying from all regions of the island and most individuals likely were present in our trapping region during the sampl ing period. 45 Table 4.1. Bald eagle (Haliaeetus leucocephalus) trapping success using the floating fish set; Langara Island, 1995; (n=148 sets). Description Rate No. Eag les Induced off Roost 4 9 % 73 Attempted to Pick up Bait 84% 61 Trapped 36% 22 Table 4.2. Bald eagle (Haliaeetus leucocephalus) trapping results, Langara Island, 1995. Age C lass Number Trapped Adult 13 Sub Adult 3 Old Immature 2 Young Immature 4 Each eagle was ass igned an age c lass based on p lumage patterns (Bortolotti 1984), we ighed with a 10 kg Peso la spring scale, and the wing chord, tail length, cu lmen, and tarsus diameter measured. Sex was determined from the bill depth and hallux length measurements (Bortolotti 1984). Up to 10 ml of blood was drawn from the brachial vein. The blood was first col lected into heparinized Vacuta iner tubes for brodifacoum residue analysis, and if there was enough, 4.5 ml was placed into a Vacuta iner tube containing buffered sod ium citrate and immediately placed on wet ice for PT evaluation (Brown 1988). A single drop of blood was placed on to a slide for a smear. The bird was banded and re leased. The banding data has been submitted to the province of British Co lumbia , Ministry of Environment, and the National Wildlife Research Centre, Hull, P Q . 46 The blood was centrifuged at 2000 x g for 15 min and the p lasma pipetted into pre-labelled 5 ml cryovials, and immediately frozen (-20° C). PT s were measured within 6-8 h of collection using the Coulter P/T Fibrinogen following the procedures for manual evaluation (Brown 1988). For control PTs , blood was drawn from bald eag les at the Orphaned Wildlife (OWL) rehabilitation centre in Delta, B.C. and in rehabilitation centres on Vancouver Island. The PT for each sample was measured three t imes. Normal human p lasma was used before and after each sample to ensure that the test was effective. The reported PT is an average of the two c losest t imes. 4.2.5 Brodifacoum Residue Analysis Frozen t issues were shipped to the National Wildlife Research Centre, Hull, P Q for preparation for analysis. T issue homogenisat ion was carried out using chemical ly c leaned instruments to avoid contamination. Extracts of t issue were ana lysed at NovaMann International, M iss i ssauga, Ontario using high pressure liquid chromatography (HPLC) with two detection systems: a) post co lumn reaction and measurement of f luorescence of brodifacoum and b) ultraviolet spectrum scann ing (for details of procedure see Hunter 1983). The compound identified as brodi facoum using f lorescence matched the UV spectrum of the brodifacoum standard. The limit of detection was 0.005 mg/kg. A s part of the quality control, NovaMann confirmed quantitative recovery of brodifacoum from liver and reported mean recoveries of 76% for fortifications at the 47 0.5 and 1.0 mg/kg levels. Fortified rat liver samples were also prepared in the field, wrapped in foil packets, placed into pre-labelled polyethylene whirl pack bags and shipped on ice. The t issues and foil were rinsed with known vo lumes of methanol and t issues homogenised. The homogenised t issues were divided into two thirds/one third by weight. The larger portion was extracted at NovaMann , the smal ler stored at N W R C . The mean quantitative recovery of brodifacoum was 54% for fortifications between 6.4 to 23 mg/kg (Table 4.3). Al l original chromatography graphical tracings from the instrumental ana lyses were carefully examined at N W R C by Dr. Bryan Wakeford to ensure both quantitative and qualitative assessment of the brodi facoum residues were correct. Table 4.3. Quantitative recovery of brodifacoum from Norway rat (Rattus norvegicus) liver fortified on Langara Island in August, 1995. T i ssue Type T issue + Fortified Brodi facoum Percent Methanol Sent Level (ug) Reported (ug) Recovery (9) Liver 10.3 0.0 0.0 Control Liver 6.8 14.22 6.94 48.8 Liver 5.3 16.25 6.97 42.9 Liver 13.7 6.97 6.22 89.2 Liver 4.7 7.83 2.79 35.6 Bald eagle p lasma was analysed for brodifacoum residue (Murphy et al. 1989) at the Department of Agriculture, State of Illinois Veter inary Diagnost ic Laboratory, U S A . The limit of detection was 0.005 ppm. An aliquot of the spike solution was sent to the lab with the samples for analysis. The quantitative recovery was 78 .1% for a concentration of 115 ug/ml spike solution. 48 4.3 Results 4.3.1 Carcass Locations of Brodifacoum Poisoned Norway Rats The majority (86.7%) of the radio-collared rats died underground in their burrows and therefore were unavai lable to avian scavengers (Table 4.4). At 8 d following initial baiting, an adult female was found dead on the beach above the high-tide line in a puddle of water, 24 h after her last known activity. Another radio col lared adult female was tracked into the forest canopy 10 d after the start of the baiting. The remains, including the head (with the radio col lar around the neck), forelegs and thoracic cage (organs removed) were found on the branch of a coastal Western Hemlock (Tsuga heterophylla) (dbh = 51 cm) about 10 m above the ground. The tree was located 36 m south of bait station M C B 1 9 , and >225 m from the north shorel ine of Langara Island. It is unclear whether it was scavenged or preyed on. There was no evidence of ejected pellets or other remains under the tree. This rat was known to have been active 4 d before, with no ev idence of further movement until the signal was tracked to the tree. The interval between the start of the intensive baiting and detected death for 15 rats ranged from 3 to 9 d (Table 4.5). There was no significant difference (P>0.05) in time to death between sex and/or age c lass. The s ignals for 2 adult males could not be detected possibly because of transmitter malfunction, or them having left the region. The collar and ear-tag of a juvenile female were found under a log, while the radio of another adult female was found because either it had sl ipped off, or the rat had been preyed upon. 49 Table 4.4. Ca r cas s locations of brodifacoum poisoned radio-tagged Norway rats (Rattus norvegicus), Langara Island, 1995. Age Above Below Predated or Unknown C l ass Ground Ground Scavenged Juveni le Female 0 4 0 1 Juveni le Ma le 0 5 0 0 Adult Female 1 2 1 a 1 Adult Male 0 2 0 2 Total 1 13 1 4 b 6.7% 86.7% 6.7% -a Found in the forest canopy. It is unclear if it was scavenged or preyed on. b Not included in percentage calculations. Table 4.5. Interval between start of poisoning and detected death of radio-tagged Norway rats (Rattus norvegicus), Langara Island, 1995. A g e and S ex C lass Mas s (g) m e a n 3 (range) mean ± s.e. Days to Death Post Start of Baiting" 3 n Juveni le Female 90 (52-145) 6 ± 1.15 4 Juveni le Male 74.4 (50-105) 8 ± 0.32 5 Adult Female 278.6 (198-353) 6 + 1.47 4 Adult Male 237.5 (163-298) 6 + 0.0 2 Mean 7 ±2 a n=5 for all age classes except Adult Male with n=4. b Days to Death= Days to last known activity + 1 day. 50 Other, non-radio tagged rats, of both age and sex c lasses were found dead above ground and col lected opportunistically by the baiting and research crews (Appendix A) . The locations of all rats found above ground is in Figure 4.1. 4.3.2 Residues in Norway Rats Found Dead Above Ground The brodifacoum residue concentrations in selected rat t issue can be found in Tab le 4.6. There was a significant interaction between age/sex c lass and t issue types (P<0.05) (Table 4.6). Ca r cass brodifacoum residue concentrat ions were similar across all age/sex c lasses . Liver concentrations were similar across all age/sex c lasses . The adult female liver brodifacoum residue concentration was significantly less than the juvenile male but similar to the adult male (Table 4.6). The absolute (mg) residue in the t issues of the rats found dead above ground are presented in Tab le 4.7. The whole carcass residue load (mg) of the rats ranged from 0.097-1.809 mg or 0.097-1.809 bait block equivalents. 51 Langara Pt. No-Name Pt. s Mcpherson Pt. Lord t Bight Explorer Bay ffm Fury Bay # Common Ravens ^ Bald Eag les • Northwestern Crows O Norway Rats t Hazardous Cove Eger ia Bay Figure 4.1. Locat ions of all ca r casses of common raven (Corvus corax) and Norway rats (Rattus norvegicus), together with sampl ing locations of bald eag les (Haliaeetus leucocephalus) and Northwestern crows (Corvus caurinus) with positive detection of brodifacoum residue, 1995. t Henslung Cove 'Hol land Pt. 52 0 l _ CO CD -K -o ° o 5 -a £ "co ^ sz co 3 CO .<•> c cn co ? l o . . C co co 0 S *S O CO CO CO CO E co j= CD 5 = c -a ^ o 3 LO CO °> £ g 2 S c E 0 w g LO o co O CD 0 ^ •3 ~° CO CO 1 co 3 co 8 g> 45 co =6 -1 2 -d to g CO 0 0 1 ° ca oo CO ICQ _CD o CD LO o o V a. CO c 0 1 _ > > -+—' c CO o 5^ 'c ral CO 2, CD 0 > CL u O £ 0 CL CO 0 E co CO 0 sz CO CO CO 2 CO o CO D) CM oo co o • CM CO co LO d> CO co 00 CO CO •55 ° ° "8 oo CO CN CM 00 CO o CM o .o CO CO co cd co i o CM CM LO a) T3 CM CO LO . O O CD OO O0 CO CO LO CM 0 CO ra LO CO CM CO CM ZJ < 0 CO E 0 u_ •4—' ZJ < CT> co CM co T LO 2 LO X CO CO oo LfT co _^ T — CM CM <D T3 CD 4.33-4 LO o O co co 4.33-4 LO CO CM LO co • CM Sri CO n CO CO Cp LO ^ CM CM O 0 CO 0 'E 0 > ZJ CO LO Table 4.7. Brodi facoum residues (mg) in Norway rats (Rattus norvegicus) found dead above ground, Langara Island, 1995; (mean+ s.e., range in brackets). A g e and Sex Ca r cas s Liver GIT Who le Body Adult Ma le 0.702±0.101 0.235+0.033 0.474+0.141 1.472±0.260 (0.495-1.001) (0.188-0.327) (0.181-0.791) (0.897-1.809) Adult Fema le 0.439±0.138 0.646+0.130 0.270±0.078 1.327±0.242 (0.101-0.757) (0.388-0.802) (0.042-0.383) (0.872-1.697) Juveni le Male 0.058±0.021 0.039±0.013 0.108±0.053 0.205±0.086 (0.037-0.100) (0.015-0.061) (0.045-0.214) (0.097-0.375) 3.3 Scavenger Identification Common ravens were identified to be the most significant avian scavenger (80%) of Norway rats (Table 4.8). Northwestern crows also were photographed at six ca rcass sites and suspected of scavenging three other rats. Table 4.8. Frequency of avian scavengers of unpoisoned Norway rat (Rattus norvegicus) carcasses, Langara Island, 1994. Spec i e s Number Percent T ime to Scaveng ing Raven 12 80 1-24 h Crow 3 20 1-14 d Total 15 100 Crows, however, were also attracted to three other Norway rat ca r casses after they were scavenged or buried by the necrophagous beetle Nicrophorus sp., or attacked by other insects (Table 4.9). Photographs of an adult bald eagle were 54 taken at a carcass that was not subsequent ly scavenged, but buried by the beetle Nicrophorus sp. Song sparrows (Melospiza melodia) were photographed at three rat ca r casses that had been attacked by carrion insects. Table 4.9. Identified scavengers of unpoisoned Norway rat (Rattus norvegicus) carcasses , Langara Island, 1994. Scavenger C lass Number Percent Av ian 15 52 Insect 5 17 Unknown 4 14 Not Scavenged 3 10 No Data 2 7 Total 29 100 Ravens scavenged carcasses within 1 h of placement for six ca r casses and between 4-24 h for the remainder. The carcasses were suspected to have been scavenged by crows anywhere from 1 h to 14 d after placement. The cameras failed to take pictures of the scavenging spec ies or took pictures for no apparent reason at 11 carcasses . 4.3.4 Effects on Predators and Scavengers 4.3.4.1 Common Ravens A total of 13 ravens were found dead during the intensive baiting campa ign between days 12 and 47 after the start of the baiting (Table 4.10 and Figure 4.1). The coast l ine between Fury Bay eastward to Iphigenia Point, yielded six po isoned ravens between days 13 and 20, or about 1.14 ravens/km of coastl ine. The east 55 coast from Hol land Point north to Mcpherson point yielded four dead ravens or 0.38 ravens/km coastl ine. Three were also found in Henslung Cove late in August, 1995. Al l were within 300 m of the shorel ine. Liver residue ana lyses confirmed all 13 ravens were exposed to brodifacoum. The level of brodifacoum residue ranged from 0.985 mg/kg to 2.522 mg/kg with a mean of 1.353 mg/kg. There were no significant differences in residue levels between mature and immature birds (two tailed t-test P>0.05) or between male and females (two tailed t-test; P>0.05) (Table 4.11). 56 "D C Zf £ S o o CO o O CO c 0 > ro i_ tz o • E E o o CD ro o E CD XZ M— o co CD ' co "O C CO CO -+—» c c o o -+—' Zl CD CO CD Z l •g 'co E i Z l • o o • •2 T3 • P • CD 2 . ro d g| •Q ro (0 CD I— -o re Q. <2 <S + S c o O i + ra or E S I'if 2 DC Q) T3 i_ O CO 2 !G c o o o O co .2> co CD CD < X CD CO •S c CO 3 o o o LL c CO > ro or CO CO ro o E co x: £• ro c o E 3 0. o CL O to Q. O CO D) ro o E co .c £• ro c o E 3 CL CO O) ro o to o E CO E CO o o ro to ro £ m CO CO ro o E CO ro c o E 3 CL CO CD ro o E co J= £< ro c o E 3 CL CO CO ro ro c o 3 CL o CO c c CO to E « o o XJ to •° 3 ro E "to ro CO i CO ro "a 3 ro o 73 o o JD CO CO CO CO CO CD CD CD ro ro ro 1 o o o E E E CO CO CO .c x: .c £• £• £• ro ro ro c c c o o o E E E 3 3 CL CL CL i 1 1 + + 1 + + 1 1 1 1 + o CM lO CO CO CO o CM o CO CO o CM CO m CM CM CM CO CO CM T— p p m OJ CO CM CM CO p CM o t— ID CO 03 O O CM CO CM O O CM CM O CM a> CO CO CO CM CM CD > o O to 3 O "2 ro N ro X CO > o o to 3 o "E ro N ro X Q ro CO £> 3 LL CO "D ro Q Q Z ro m 3 ro m £• 3 LL LL CO > o o o "E ro N CO X CM CO 00 05 O) T— CM CM CM CM CM CM CM CO >» >, >, >, >, >% 3 3 3 3 3 3 3 3 —> 3 ~3 —> ~3 ~3 -3 -3 ro m I .a Q to 3 CD 3 < ro CO co CD LU CO to 3 CD 3 < LO CD CO CD > > > O o O O O O CD CD CD C C C 3 3 _3 to (0 to c c c CO CO CO X X X CO CO to CM CM CM •4-* •»-> to to (0 3 3 3 CD CO CO 3 3 3 < < < CM CO to D) -9 T3 2 to CU a) *s <u to > £ II .£ 1 0 T3 T3" "> ° CD ™ M to •* 0) ,_- si 'ra c u. -o o o ,_ 0- ro T3 CO U | 8 3 co Us 1 o o o O ro cu o <n o 3^ ** ro <o 8 § $ 8 0) . to c Q) 'F "5 C C (U LO Table 4.11. Common raven {Corvus corax) liver brodifacoum residue levels (mg/kg), Langara Island, 1995 (geometric mean; 9 5 % conf idence interval in brackets). e S ex Immature (n=6) Mature (n=7) Male (n=6) Female (n=4) 1.18 (0.98-1.41) 1.45 (1.07-1.91) 1.24 (1.03-1.48) 1.24 (-0.82-28.05) Necropsy revealed that 6 9 % of the ravens had died from severe pulmonary hemorrhaging and the remainder of intramuscular or intracoelemic hemorrhaging (Table 4.12). One death could not be determined due to autolysis, however, the liver brodifacoum concentration was similar to the others. Table 4.12. Pr imary sites of hemorrhage in common ravens (Corvus corax) found dead, Langara Island, 1995. Pr imary Site of Percent No. Hemorrhage Pulmonary 69% 9 Breast Musc le 15% 2 Intracoelemic 8% 1 Undetermined 8% 1 Source of Brodifacoum Poisoning to Ravens Only seven raven g izzards contained sufficient material for protein electrophoresis. A protein band matching rat or muskrat haemoglobin was seen in one. No rat serum albumin was observed. No protein bands matching shrew controls were observed and no bands from bait or snai ls. Four samples showed protein matching raven albumin. 58 Table 4.13. Food remains in the g izzards and intestines of 13 common ravens (Corvus corax) found dead, Langara Island, 1995. Contents Frequency No. (%) Rat Hair 38 5 Bait 31 4 a Invertebrate Marine 46 6 Other 15 2 Vegetat ion Marine 46 6 Other 46 6 Vertebrate Bone 8 1 Unidentified 46 5 a- Hair characteristic of Norway rat also found in two. Five (38%) of the gizzard and intestinal contents contained hair characterist ic of Norway rats, including the contents from the raven which showed positive for rat or muskrat haemoglobin in protein electrophoresis (Table 4.13). The gizzard contents of one raven contained bait block fragments mixed with 11 unidentified avian bone fragments (1.09 ± 0.12 cm long x 0.46 ± 0.05 cm wide (mean+s.e.)). No rat hair was found in the gizzard or intestinal contents. Three regurgitated raven pellets were found in Eger ia Bay on day 35 post start of baiting (16.8 ± 1.9 g, 4.45 ± 0.31 cm long, 2.66 ± 0.08 cm wide (mean t s.e.)). Al l were situated on logs, two on the beach above the high tide line, and one under the forest canopy. The blue brodifacoum bait predominated in the pellets, although hairs characteristic of Norway rat were also found in each . 59 Raven Activity -1996 Five nests were visited, however, none were active (Table 4.14). A nest in Dibrell Bay was empty and the bark and tops of branches under the nest were coated with blue tinted faecal matter. The remains of a scavenged raven were found nearby. Prey remains under the nest included chitons and 12 regurgitated pellets consist ing primarily of bait and seven distinct piles (1- >100 pieces) of bait crumbs. Stretched pieces of plastic bag that had wrapped the bait were found in 58% of the pellets. Three more regurgitated pellets consist ing of the blue bait were found between No-Name Point and Dibrell Bay, while searching for raven nests. Table 4.14. Inactive common raven (Corvus corax) nests, Langara Island, 1996. Nest # Geograph ic Yea r of Last Tree Spec ies Tree Nest Locat ion Known Activity d b h a H A G b 1 Mcpherson Point 1995 Sitka Spruce 18 m 2 Mcpherson Point - Western Hemlock 10 m 3 Mcpherson Point 1994 Western Hemlock 7 m 4 Dibrell Bay - Sitka Spruce 1.16 m 15 m 5 Hazardous Cove 1995 Western 0.73 m Hemlock 7.30 m a dbh: Diameter at breast height, b HAG: Height above ground. 60 The prey remains under the nest in Hazardous Cove consisted of snai ls (5 Haplotrema sp. and 1 Vespericola sp.), two limpet shel ls, black-tailed deer hair, eight chitons and the skull of an adult Norway rat. P last ic bags were found outside 54 of 107 bait stations investigated. Tear patterns characteristic of the raven positive controls were found on 9 8 % of the bags. The imprint of a beak and/or bill tip of ravens were found on the a luminum trays from 12 stations. The imprints were within 1.7±0.26 (mean ± s.e.) cm of the tray edge which faced either opening of the stations. Bait crumbs were found outside stations or on top of nearby logs at nine bait stations. The remains of seven more ravens were found or reported including a fresh, dead raven at the top of the beach in Hazardous Cove (Table 4.15). Necropsy results confirmed its death was due to severe, bilateral pulmonary hemorrhaging. Four pairs of ravens were observed flying along the north end of the island between Langara Lightstation and Mcpherson Point, in Eger ia Bay, Hazardous Cove , and at Lord Bight on the west coast. This indicates that some ravens were still al ive on Langara Island in 1996. 61 CO o O co' . CO CO CO CO T -a ? • i5 5 2 •8 * X I C (0 CO I— —I E E o o CO 1— CD CL co" u 0 0 SZ o CO CO c 'co E 2 CO -4—1 0 0 CO c 'o CL l _ CO T3 CO CL < ZJ < CO 0 £Z o CD L— 0 sz -+—' CO 0 LL •3 0 CD £Z 0 > CO o CO CO CD I -Q b CO CM E E o o £2 0 CL * f o LU -j CO l _ 0 sz CO 0 c "o CL 0 E co Z 6 z CO CO o sz -4—1 ZJ < CO c 'co E 0 I— TO -4—• 0 0 J* CO 0 > o o CD cz _ZJ CO c 0 CO ZJ < ZJ < CO c 'co E 2 CO -I—» 0 0 CO CO 0 CO CO TO " i _ 0 CD LU CO 0 > o O CO ZJ 0 "2 CO N CO 1 CO LO CO E E o o 0 CL i_T 0 CO 'co CO c 'co E 2 CO -*—' 0 0 JXL CO CO CO CO ' l _ 0 CD LU co ZJ CD ZJ < CM CO 4.3.4.2 Northwestern Crows A summary of information on the crows collected is in Append ix Tab le 4-3. Brodi facoum was detected in one pooled sample of crow livers col lected from Lucy Island, 12 d after the start of the baiting campaign in 1995 (Table 4.16). A crow was found dead in Hens lung Cove on August 7, 1995 but no brodifacoum was detected. Trace amounts of brodifacoum (0.048 mg/kg) were detected in one crow col lected in May 1995 before the intensive baiting began on Langara Island. Table 4.16. Brodi facoum residue levels in livers of Northwestern crows (Corvus caurinus), Langara Island, 1995. Pool No. Crow #'s Brodi facoum (mg/kg) 1 a 1 0.048 2 a 15, 16 N D C 3 2, 3 , 4 , 5 ND 4 6 , 7 , 8 0.019 5 9, 10, 11 ND 6 12, 13, 14 ND 7 15, 16 ND 8 b 17 ND 6 18 ND 9 19 , 20 , 21 ND 10 22, 23, 25 ND 11 24 L A d a Collected in May 1995. b Found dead in Henslung Cove. c None detected. Detection limit <0.01 mg/kg. d Lost in analysis. 4.3.4.3 Bald Eagles A total of 22 bald eagles were trapped; two before the baiting began, and 20 over the course of the baiting for a 36% trapping success rate. Another eagle was 63 rescued from a surge channel at No-Name Point 3 d before the eradication program. Mainly adult birds were targeted although four young of the year were also trapped. The overal l sex ratio was 13 males: 10 females. A detailed summary of eag les trapped is in Append ix Table 4-2. Brodi facoum was detected in the p lasma of three individuals (15%). Each was from a different age c lass and two were female. The greatest p lasma residue was detected in the subadult trapped at Cohoe Point (Table 4.17 and Figure 4.1). Table 4.17. Ba ld eagle (Haliaeetus leucocephalus) p lasma brodifacoum residues and prothrombin t imes, Langara Island, 1995. A g e Brod i facoum P T No . Date Locat ion S e x C l a s s 3 Res i due (mg/kg) (sees) 1 Ju ly 5 Margaret Point M A N D -2 Ju ly 6 Dibrel l Bay F A N D -3 Ju ly 7 No -Name Point M S A N D -4 Ju ly 18 Exp lorer Bay M A N D -5 Ju ly 19 Mcphe r son Point M A N D -6 Ju ly 23 Iphigenia Point M O l N D -7 Ju ly 31 Iphigenia Point F A N D -8 Augus t 1 No -Name Point M Y l 0.041 -9 Augus t 1 No -Name Point F S A N D -10 Augus t 3 No -Name Point M A N D 147 11 Augus t 4 No -Name Point F A N D 261 12 Augus t 10 Eger ia Bay F A 0.037 -13 Augus t 10 C o h o e Point M A N D 177 14 Augus t 10 C o h o e Point M A N D 201 15 Augus t 10 Dibrel l Bay F A N D 234 16 Augus t 11 No -Name Point M S A N D 224 17 Augus t 12 And rews Point M A N D 151 18 Augus t 13 C o h o e Point F S A 1.74 221 19 Augus t 14 And rews Point M Y l N D 284 20 Augus t 15 Margaret Point F Y l N D 185 21 Augus t 25 Iphigenia Point F O l N D 122 22 Augus t 26 Dadens F A N D 226 23 Augus t 26 Hens lung C o v e M Y l N D 122 M e a n ± s.e. 197 ± 14 a A= Adult; S A = Sub-Adult; OI= O ld Immature; Y l= Y oung Immature (Bortolotti 1984) 64 PTs were not performed for the adult and young immature bird exposed to brodifacoum either due to lack of blood col lected, or because lipids in the p lasma made them unsuitable for testing. The subadult eagle P T was 221 s (Table 4.17). The mean PT for all eagles was 197 s which was significantly longer (one tailed t-test P<0.05) than the mean control PT of 125 s (Table 4.18). Table 4.18. Bald eagle (Haliaeetus leucocephalus) control prothrombin times. Number Sex A g e a PT (sees) 1 M YI 70 2 F YI 177 3 M Ol 151 4 M YI 125 5 M YI 59 6 F A 149 7 F Ol 108 8 F YI 146 9 F YI 145 10 M YI 145 11 M YI 97 mean + s.e. 125 ± 11 a Yl= Young Immature; OI= Old Immature; A= Adult 65 4.4 Discussion 4.4.1 Carcass Locations of Brodifacoum Poisoned Norway Rats A total of 13.4% of radio collared Norway rats died above ground and all were adult females. This is likely an underestimate of the actual frequency of death above ground, because other non-radio collared rats found dead above ground included all age and sex c lasses. The location of death does not appear to be age or sex related. In 1994, one of three radio-collared rats died above ground after the poisoning operation on Lucy Island (Howald 1995). These results are in apparent contrast to New Zea land operations where all 16 radio-collared Norway rats on Ulva Island died in their burrows (Taylor 1993) and no poisoned rats were found on the surface of Hawea Island (Taylor and Thomas 1989). Further, only four Norway rats were found above ground on Breaksea Island in New Zealand after similar brodifacoum baiting for rats (Taylor and Thomas 1993). However, there was no indication of a quantified search method in these studies. During anticoagulant poisoning operations on farms in England, a small proportion (%) of the Norway rat population was found to have died above ground, the majority died below ground (Harrison et al. 1988; Fenn et al. 1987). In laboratory enclosure trials with wild Norway rats, 67% were reported to have died in open areas after receiving a lethal dose of brodifacoum (Cox and Smith 1992). Similarly, Gemmeke (1990) showed that captive anticoagulant poisoned rodents died above ground as often as below. More work is required in this area to accurately determine 66 the proportion of rats that die above ground, thus making themselves available to scavengers. The remains of the adult female found in the tree confirms that some rats were either scavenged or preyed on. The scavenger or predator could not be identified but the location of the tree 225 m inland suggests that a raven was likely responsible. Bald eagles were rarely observed under the forest canopy after the ancient murrelets had left the colony in mid-June. Ravens, however, were regularly observed in the forest and were identified as the most significant scavenger of rats. The mean time to death for the radio-collared rats was 7 d after the start of the baiting program. Although the precise length of the latent period was not known, it does indicate that the majority of rats were dead and available to scavengers a week after the baiting program began. For predators, the greatest opportunity to catch a live but toxic rat, was within the first week of the baiting campaign. 4.4.2 Residues in Norway Rats Found Dead Above Ground The presence of brodifacoum in the carcasses of poisoned rats found above ground poses a secondary poisoning hazard to the identified scavengers. The liver contained the greatest concentration of brodifacoum as compared to the carcass, but was not significantly greater than the GIT. This preferential storage in the liver is consistent with other anticoagulants such as Warfarin, bromadiolone and f locoumafen (Huckle et al. 1989a; Huckle et al. 1989b; Lechevin and Vigie 1992; Newton et al. 1990). The liver is the target organ for anticoagulant rodenticides where they act to 67 prevent the reduction of the vitamin K epoxide by inhibiting the epoxide reductase enzyme (Mount 1988). However, the liver constitutes only 4 .1% by weight of the rat, while brodifacoum is present throughout the carcass. Thus, adult rats present more of a hazard than smaller rats because of their weight which allows for greater brodifacoum accumulation. The GIT was the t issue that resulted in the significant interaction effect between age/sex c lass and t issue type. This was likely a result of the presence of unassimi lated bait found in the alimentary tracts of some individuals. On dissect ion, some stomachs were found to be packed with bait and the intestines were tinted with the colour of the bait. The juvenile males contained the greatest concentration of brodifacoum in the GIT suggest ing that they preferred and continued to feed on the bait even after hemorrhaging had begun. Adult females appeared to contain the least amount of brodifacoum concentration and this may reflect the demands placed on her, such as providing for the pups by bringing back bait to the burrows. This could account for the high concentration in the juvenile GITs. The mean whole body brodifacoum residue concentrations found in the rats on Langara seemed to be the highest yet reported and appears to reflect the saturation baiting strategy used. Dubock (1984) reported a mean carcass residue level of 3.2 mg/kg in rats after saturation baiting with brodifacoum, although 8% contained a concentration above 10 mg/kg. Merson et al. (1984) live trapped, euthanised and ana lysed the whole body brodifacoum burdens of meadow voles (Microtus pennsylvanicus) and found them to range from 2.07+ 0.17 mg/kg (mean+s.e) to 68 4.07± 0.20 mg/kg after aerial application of bait. In laboratory trials, brodifacoum concentrat ions of voles fed 10 mg/kg brodifacoum ranged from a mean of 0.40 to 0.53 mg/kg, while voles fed 50 mg/kg brodifacoum bait ranged from a mean of 2.17 to 5.21 mg/kg (Kaukeinen 1982). In general, carcass residue concentrations are positively related to the rate of anticoagulant application (Merson et al. 1984; Dubock 1984). The distribution of brodifacoum across different t issues may provide a degree of secondary poisoning protection, if there is rejection of any of those t issues. For example, golden eagles were observed to reject the stomach and entrails of strychnine poisoned Richardson ground squirrels (Spermophilus richardsoni) resulting in no detectable strychnine poisoning impact to any eagles (Graham 1977). If a brodifacoum poisoned rat is eviscerated and the GIT not consumed, between approximately 30-50% of the total brodifacoum residue is thereby avoided. However, there was wide variability between each age and sex c lass in t issue retention of brodifacoum. This variation may be due to the small sample s ize and further work is required. 4.4.3 Common Ravens Common ravens were the most significantly impacted spec ies in the study. Although no population estimates were made, the 20 ravens found dead between July 1995 and August 1996 were likely a minimum number affected. Carcass searching studies have had low success rates, due to the efficiency with which scavengers and 69 predators removed the carcasses, as well as the low efficiency of human searchers (Linz et al. 1991; Brown et al. 1988; Mineau and Coll ins 1988; Ba lcomb 1986; Stutzenbaker 1984). Therefore, the impact on the raven population can only be estimated. On Saltspring Island, Brenchley (1985) estimated the minimum raven nesting density to be 26/100 k m 2 . A high estimate relative to Europe where 17 nests/100 k m 2 was the highest estimated density (Newton et al. 1982). Assuming Langara could sustain similar densities, there may have been five to eight territorial, nesting pairs with between three to seven fledglings each (Ehrlich et al. 1988) before the baiting began. Therefore, the pre-baiting common raven population may have been from 20 to 72 individuals, and the impact of the poisoning operation on the local raven population appears to have been heavy. However, the presence of four pairs in 1996 indicates that some individuals survived, or that empty territories were taken over by immigrants from nearby Graham Island. Necropsy Results and Liver Residue Analyses The necropsies confirmed that all ravens died of acute, multifocal hemorrhage, symptomatic of anticoagulant poisoning (M. Mcadie, pers. comm.). These results are in accordance with other studies investigating anticoagulant poisoning in avian predators (Mendenhall and Pank 1980; Newton et al. 1990). The body condition scores indicate that none of the birds were emaciated or in poor condition, and there 70 was no evidence of d isease that may have predisposed the ravens to the effects of brodifacoum. The role of brodifacoum poisoning could not be determined when only feathers and/or skeletal remains were found. Some mortality may have occurred over the winter, independent of brodifacoum use. However, the locations and evidence where individuals were found suggests that brodifacoum poisoning was the ultimate cause of death, even though predation may have been the proximate cause or they were scavenged after death. A s brodifacoum binding in the liver is a saturable process resulting in a steep dose response (Thijssen and Baars 1989; Godfrey 1985), and death from anticoagulants is delayed allowing for some metabolism and excretion of the ingested dose (Godfrey 1985), the relationship between the concentration measured and the initial exposed dose is complex and difficult to interpret. Godfrey et al. (1985) found that liver levels are unsuitable for quantifying exposure to brodifacoum because of the lack of correlation between liver residue concentration and dosing levels. However, it can be concluded that the positive analysis of brodifacoum in the liver together with the necropsy results are symptomatic of brodifacoum poisoning. Source of Brodifacoum Poisoning to Ravens Some raven mortality from scavenging dead rats was expected, however, primary poisoning from raiding bait stations was not. The first indication that primary poisoning contributed to raven mortality came after the bait spill at the Langara Fishing 71 Lodge in Henslung Cove. Two to three 20 I buckets containing old bait blocks awaiting incineration, were left open and unattended overnight. The following morning, the fishing lodge manager reported a "sea of black" ravens picking up bait blocks in their bills and flying off with them. Over the next 4 d, three ravens were found dead and two were observed falling off their roost and picked up dead. This may have been an isolated primary poisoning event because it was not uncommon to have piles of garbage accumulate before incineration, and to observe ravens work open holes in bags and boxes. The presence of the bait buckets with open lids presented an easy opportunity for ravens to investigate a new food source. The attraction to the bait indicates the palatability of this toxic food source to ravens, however, it is not known if ravens were responsible for raiding bait stations during the intensive baiting in 1995. There were no reports of ravens raiding bait stations during the initial campaign and stations were not investigated as a source of poisoning. If ravens were removing bait from stations, it would have been assumed rats were responsible. The beak marks on the aluminum trays, regurgitated pellets of bait containing bits of plastic bag, and torn bags found outside bait stations, confirmed that ravens raided bait stations over the winter of 1995/1996. The systematic checking of bait stations by operators might have been observed by ravens and led to their investigating the stations, finding the bait blocks, and being primarily poisoned. Ravens preyed on pinyon jay (Gymnorhinus cyanocephalus) eggs and young within 24 h after researchers had climbed trees with nests (Marzluff and Balda 1992). 72 Similarly, in the Okanagan Val ley of British Columbia, ravens have been observed systematical ly checking orchards for songbird nests, perhaps imitating researchers studying them (A. Preston pers. comm.). In May 1996, eight bait stations, including four all in order and along a line, that were checked and re-armed 2 d previously, had the bait removed. The design of the stations and how they were staked down, allowed ravens to employ one of three tactics to gain access to the bait in them: reaching into the station and pulling the aluminum tray and bagged bait blocks out, lifting one end of the station so that the bait rolled c loser to either opening, or uprooting the stations from the stakes so the bait rolled out. In New Zealand, western wekas and keas (Nestor notabilis) have been observed reaching into the stations and pulling bait out (Eason and Spurr 1995; Taylor and Thomas 1993). However, the action of the ravens on Langara is apparently the first reported case of non-target species completely disrupting bait stations to gain access to the bait. Clearly, the bait stations were inadequate to exclude common ravens and need to be modified for future rat eradication programs. The primary and secondary poisoning appears to have had a large impact on the local common raven population on Langara Island. In New Zealand, the entire population of western weka was exterminated from Tawhitinui Island after feeding on brodifacoum bait intended for ship rats, and by eating dead or dying rats (Taylor 1984 in Eason and Spurr 1995). Similarly, 80-90% of Stewart Island weka were primarily and secondari ly poisoned (Eason and Spurr 1995). Both wekas and common ravens 73 share an aggressive and inquisitive behaviour and opportunistic, omnivorous diet (Falla et al. 1983; Ehrlich et al. 1988) which contributed to their decline in population. The gizzard analysis confirmed that ravens were not consuming exclusively either rats or bait during the initial campaign but the content analysis is not a sensitive enough indicator to determine the source of poisoning. No rat hair, rat protein or bait were detected in six dead ravens in 1995 probably because of the delayed time to death after ingestion of brodifacoum allowing for evacuation of the GIT contents (Godfrey 1985). Captive barn owls feeding on brodifacoum poisoned rats died between 5-6 d after the last poisoned rat was consumed (Mendenhal l and Pank 1980). It is possible, however, that brodifacoum exposure may have originated from another source such as invertebrates found feeding on the bait (See Chapter 5). The presence of bait crumbs along with Norway rat hairs may indicate that ravens were both primarily and secondarily exposed to brodifacoum, or that the bait crumbs originated from the rats. The latter seems more likely because on dissection, rat stomachs and intestines were laden with bait crumbs. However, the presence of bait crumbs without the detection of Norway rat hair, does not preclude the possibility of secondary poisoning. For ravens 11 and 12, where only bait crumbs were detected in their gut contents, it seems probable that primary poisoning was the cause because they were found within 4 d after the bait spill in Henslung Cove. Secondary Poisoning and Toxicoiogicai Significance Rats were confirmed as a source of brodifacoum to ravens, however, it is unclear if one or more rats were required to impair hemostasis and cause death. The 74 L D 5 0 , the single acute, oral dose required to cause death to 50% of a population, is used as an index of toxicity and demonstrates a pharmacological risk of brodifacoum. The L D 5 0 of brodifacoum for the raven is unknown, however, I calculated a value that offers 9 5 % spec ies protection, with 95% and 50% confidence, from 10 published brodifacoum L D 5 0 values representing nine major families of birds (Aldenberg and S lob 1993). In other words, the L D 5 0 with 95% and 50% confidence estimate for 95% spec ies protection based on published values (Table 4.19). For any unknown spec ies the L D 5 0 would be above 0.56 mg/kg (50% confidence) or 0.105 (95% confidence). The closest related spec ies to the raven with published L D 5 0 values are from the Order Passeri formes, specifically the house sparrow (Passer domesticus) with an L D 5 C >6 mg/kg, blackbird (Turdus merula) and hedge sparrow (Prunella modularis) >3 mg/kg (Godfrey 1985). These values represent the highest dose administered. Caution must be used when estimating hazard because the c lose taxonomic relation cannot be consistently used to predict the sensitivity of birds to pesticides (Hill 1994). Therefore, the L D 5 0 range for ravens may be between 0.56 and >6 mg/kg. With a mean weight of 1.19 kg for all ravens found dead over the summer of 1995, as little as 0.69 mg to > 7.14 mg brodifacoum could result in a 50% chance of lethal hemorrhaging. A single adult rat, with a mean residue level of 1.4 mg, would result in the equivalent dose of 1.17 mg/kg if consumed whole. Captive common ravens fed diphacinone poisoned adult rats consumed everything except the skin and bones (J. Marzluff, pers. comm.). However, small rats were eaten whole and a pellet was produced. The regurgitation of a pellet would decrease the amount of brodifacoum 75 X ingested and possibly, decrease the risk of hemorrhaging. Barn owls feeding on brodifacoum poisoned mice and producing pellets, reduced the amount of consumed rodenticide by an average of 25%, thus lowering their risk of hemorrhage (Gray et al. 1994). Assuming ravens regurgitated a pellet containing 25% brodifacoum, this would decrease the single brodifacoum dose to 0.88 mg/kg. This would still be within the estimated L D 5 0 range, thus a single adult rat may pose a secondary poisoning hazard to ravens. Primary Poisoning and Toxicologicai Significance The toxicologicai significance of bait blocks to ravens is difficult to accurately interpret because ravens regurgitate a pellet (due to the high, indigestible,- paraffin wax content) that likely contains brodifacoum. Each block contains 1 mg brodifacoum, enough for a raven to have a 50% chance of haemorrhaging, if all the brodifacoum was ingested. Obviously, the chance of lethal poisoning increases with the number of blocks that are consumed, and as each station was armed with three bait blocks over the winter months, each raven was exposed to as much as 3 mg brodifacoum at each bait station it raided. It seems probable that any one raven upon discovery of the easy food source found in a bait station, continued to raid more than one station. 76 Table. 4.19. Spec ies and L D 5 0 values used for calculating the value that offers 9 5 % bird spec ies protection with 95% and 50% confidence limits. Common Name Latin Name L D 5 0 (mg/kg) Reference Black-Backed Gull Larus dominicanus <0.75 Godfrey 1985 Pukeko P. porphyrio melanotus 0.95 Godfrey 1985 California Quail Callipepla californica 3.3 Godfrey 1985 Mallard Du c k a Anas platyrhynchos 2.3 Godfrey 1985; Ross et al. 1978 Harrier Hawk Circus approximans 10.0 Godfrey 1985 Ring Necked Phasianus colchicus 10.0 Godfrey 1985 Pheasant Paradise Shelduck Tadorna variegata >20.0 Godfrey 1985 Chicken a Gallus gallus 3.15 Ross et al. 1977; Taylor 1993 Japanese Quail Coturnix japonica 11.6 Ross eta l . 1976 a Mean of LD5 0 values in literature. Conf idence limits for hazardous concentrations based on logistically distributed brodifacoum L D 5 0 value calculated following Aldenberg and S lob (1993). Two rules were followed as employed at the National Wildlife Research Centre (P. Mineau, pers. comm.): 1. Include known absolute L D 5 0 values. Do not use < or > L D 5 0 values if they are within the range of the absolute values. 2. If < or > LD50 values fall outside of absolute L D 5 0 range, include and new range has been establ ished. The 9 5 % and 50% confidence limit for brodifacoum L D 5 0 data is 0.11 mg/kg and 0.56 mg/kg. 77 4.4.4 Northwestern Crows Brodifacoum was detected in low levels in the sampled crows. The crow collected in May 1995 was exposed to brodifacoum from Lucy Island because brodifacoum was not available on Langara Island until July 1995. It is unknown if exposure occurred during the baiting on Lucy Island in 1994, or if exposure occurred after the bait was removed from the stations in August 1994. The biological half-life of brodifacoum has been estimated to range between 120-220 d (Godfrey 1985; Parmar et al. 1987). With such a range, exposure may have occurred during the baiting on Lucy Island. Brodifacoum was still present and available on Lucy Island 9 months after the removal of the bait from the stations. In May 1995, four old bait blocks (9.6 g) were found under a log on the west side of Lucy Island and contained 10.99 mg/kg brodifacoum. Snai ls found near the bait were collected and contained a concentration of 0.910 mg/kg brodifacoum. Therefore, it is possible that crows could have been exposed to brodifacoum from preying on invertebrates, that had fed on old bait, during the winter of 1994/1995. Inspection of the adult rat carcasses visited by crows, but not scavenged, revealed patches of missing hair. The rats had also been moved from their original placement. The photographs revealed that crows were grabbing the carcasses by the tails, legs and flanks. The crows were apparently unable to fly off with the larger rats and were unable to break the skin of fresh, dead rats. Thus, brodifacoum exposure from scavenging rats was minimised by the crow's inability to carry whole rats to their 78 perches for consumption, and their inability to break the skin to gain access to the contaminated flesh. 4.4.5 Bald Eagles Bald Eag les were confirmed to have been exposed to brodifacoum but no mortality was detected during the intensive baiting in 1995. Similarly, no evidence of lethal poisoning to New Zealand falcons (Falco novaeseelandiae) or moreporks (Ninox novaeseelandiae), the main avian predators at risk, was noted on Breaksea Island in New Zea land, though they did prey on rats (Taylor and Thomas 1993). Further, comparable numbers of southern great skuas (Catharacta lonnbergi) and New Zea land Falcons were seen before and after a similar baiting operation on Hawea Island in New Zealand (Taylor and Thomas 1989), but, these numbers could have been maintained by immigration. Eag les were not attracted to rat carcasses put out for them, but may have preyed on rats during the latent period. Eagles are known to take rodents and other mammals associated with a body of water, such as muskrats (Ondatra zibethicus) (Knight et al. 1991; Watson et al. 1990). On the San Juan Islands, rabbits were identified as the eagle's main prey item (Retfalvi 1970). Eagles spend much of their time perching on trees overlooking the water and shoreline (Gerrard and Bortolotti 1988) and may have opportunistically preyed on a rat if it displayed symptoms of anticoagulant poisoning near or in the intertidal zone. Rats were likely available for diurnal eagles because 35% of rats were live-trapped during daylight h and were 79 known to utilise the beaches, because the contents of a rat burrow on Lucy Island contained a carapace and legs of a shorecrab (Howald 1995). One of the radio-collared rats (see above) and other rats were found dead on or near the beach, and two rats were found to have died sometime in the afternoon, including one on the beach edge above the winter high tide line. The apparent low brodifacoum exposure of the bald eagle population was likely due to the highly productive marine environment around Langara Island which is the major, if not exclusive, foraging area. Bald eagles were commonly observed fishing for Pacif ic sandlance (Ammodytes hexapterus), Pacific herring (Clupea harengus) and salmon (Oncorhynchus spp.). In 1994, over 100 eagles were observed fishing off the west coast of Langara Island. Knight et al. (1990) reported that, for coastal nesting bald eagles, seabirds are the predominant prey item. The spatial separation of the foraging and baiting areas reduced the risk of poisoning to bald eagles. Likewise, the potential for barn owl mortality with brodifacoum baiting around farm buildings appeared to be low because they preyed on rodents in grasslands away from the buildings (Hegdal and Blaskiewicz 1984). The remains of five eagles (4 juveniles and 1 adult) were found in 1996. However, the carcasses were too autolytic to determine cause of death. The talons exhibited multiple focal areas of reddish discoloration, possibly representing excoriation of sca les or subcutaneous hemorrhage (M. Mcadie, pers. comm.) an indication of possible brodifacoum exposure. However, the talons were too desiccated to distinguish between lesions and artefacts of post mortem change. The 80 lesions were similar to those seen in thermal burns from electrocution (M. Mcadie, pers. comm.) but there are no electrical power lines on or near Langara Island. Although eagle mortality was detected 10 months after the initiation of the baiting program, the role of brodifacoum poisoning could not be ascertained. Brodifacoum may have been responsible for some mortality because of the widespread primary poisoning of ravens which, in turn, could have been preyed on or scavenged by eagles. A dead bald eagle was reported next to the remains of a scavenged raven near the lightstation (J. Schweers, pers. comm.). At least three other remains of ravens were scavenged, however, the scavengers could not be identified. Food items collected under bald eagle nests in the San Juan Islands included common ravens and Northwestern crows (Knight et al. 1990). It is important to note that winter mortality of eagles would likely have occurred independent of brodifacoum use. Overwinter mortality rates for eagles in A laska ranged from 5% to 29% (Bowman et al. 1995). It was estimated that 46% of first-year eagles, and 9% of adults died over the winter in Maine (McCol lough 1982; 1986 in Gerrard and Bortolotti 1988). On Langara Island, it was not uncommon to find bones and/or carcasses of bald eagles before baiting began. The effect of brodifacoum poisoning on the breeding population appears to have been negligible. Nesting success declined from 57% in 1995 to 3 5 % in 1996; but there was a decline from 39% to 10% nesting success in South Moresby National Park in the Queen Charlotte archipelago for the corresponding period (J. Elliott, pers. comm.). 81 Toxicological Significance of Residue Levels and Prothrombin Times No brodifacoum L D 5 0 data are available for bald eagles. The above estimate of 0.56 mg/kg is likely low for bald eagles because the mean L D 5 0 for two more closely related species, the harrier hawk (Circus approximans) -10.0 mg/kg) and the Amer ican kestrel (Falco sparverius) -8.20 mg/kg (Taylor 1993) is much higher. Thus, using a mean weight of 4.5 kg (Range: 3.6 - 5.6 kg) for all bald eagles sampled, the dose required for a 50% chance of lethal hemorrhaging may be as low as 2.52 mg ( L D 5 0 : 0.56 mg/kg) but is more likely around 39 mg brodifacoum. This is equivalent to 1.4 -21.6 adult male rats each with a brodifacoum residue loading of 1.809 mg. Thus, the risk of an eagle hemorrhaging after ingesting a single rat appears to be low. The 4.3 kg sub-adult eagle trapped on day 33 contained the highest concentration of brodifacoum (1.74 mg/l plasma). Assuming the p lasma volume (PV) in bald eag les is similar to that in red-tail hawks (mean of 3.5 ml/100g body weight, Bond and Gilbert 1958), the sub-adult eagle would have a P V of 0.151 litres. This yields a total p lasma brodifacoum loading of 0.262 mg at the time when the blood sample was drawn. The immature and female bald eagles had total p lasma brodifacoum loadings of 0.0052 mg and 0.0073 mg respectively. The plasma brodifacoum levels, however, are a poor indicator of the initial dose (Appendix 1; Huckle et al. 1989b). The plasma brodifacoum loading of Japanese quail (Coturnix japonica) 24 h post dosing with brodifacoum at 1.4, 0.7 and 0.35 mg/kg, represented only 1.16%, 0.32% and 0.19% of the original dose respectively (Appendix 1). Assuming that quail are a representative model and that the eagles 82 were trapped and a blood sample drawn around 24 h after exposure, a rough estimate of the original dose can be made. The plasma concentration of the sub-adult eagle (1.74 ppm) was greater than the p lasma concentration of the Japanese quail at 24 h post dosing, therefore, it was likely exposed to more than 1.4 mg/kg brodifacoum or more than 6 mg brodifacoum, which is equivalent to more than 3.3 rats (1.809 mg brodifacoum/rat). Again, a single rat was not likely the source of brodifacoum, but more likely having come from consuming a raven or several rats. The PT is used to screen for deficiencies in the vitamin-K dependent clotting factors. The use of a mammalian thromboplastin for avian PT determinations results in wide variation within and among samples (Appendix A; Griminger 1986). A s well, a sub-lethal exposure to brodifacoum would likely not have been detected because a 30% concentration of these factors may provide a normal PT (Brown 1988; Hoffman et al. 1988). A s well, a markedly increased PT would likely have corresponded to a positive detection of brodifacoum residue in the plasma. Therefore, it cannot be concluded that the bald eagles were at risk of hemorrhaging because a significantly longer PT was detected. Savarie et al. (1979) found golden eagles (Aquila chrysaetos) experienced increases in the PT from the control time of 23 s to 900 s (39 times longer than the control time) without mortality after feeding on diphacinone contaminated meat. Some of those golden eagles, however, did experience hemorrhaging. The effectiveness of the mammalian thromboplastin to detect a significant increase in PT was investigated (Appendix A). The thromboplastin kit was able to 83 detect a significant increase in the PT of Japanese quail dosed with brodifacoum. The quail had a significant longer PT at 2.5 times the mean control time at 72 h post dose. Assuming quail are an effective model, no eagles were clinically at risk of hemorrhaging, i.e., all were less than 2.5 times the control time, as detected by the Coulter PT fibrinogen kit. The risk of hemorrhaging to the exposed individuals appeared to be low. 4.5 Conclusions Secondary and primary poisoning studies can be divided into three levels as defined by Colv in and Hegdal (1987):' hazard to individuals, short term population effects and long-term population effects. This study was designed to address the hazard to scavengers and predators at an individual level. A secondary poisoning hazard to avian predators and scavengers does exist from the use of brodifacoum to eradicate rats from seabird colonies along the British Co lumbia coast. Norway rats were shown to die above ground and be available for avian scavengers and predators. Common ravens were the spec ies most significantly impacted as a result of both primary and secondary poisoning. The population effect remains unqua l i f i e d , however, breeding pairs were observed alive one year after the baiting program was initiated. There was a low risk of brodifacoum poisoning to bald eagles from dead toxic rats, although plasma brodifacoum levels established that brodifacoum was present in the food chain. 84 Appendix Table 4-1. Dates, locations, weight, age and sex of Norway rats (Rattus norvegicus) found dead above ground, Langara Island, 1995. Rat # Date Location Found Weight Age Class Sex Reported/ (g) Turned In 1 - Mcpherson Pt. - A F 2 - Mcpherson Pt. - J M 3 - Mcpherson Pt. - A F 4 - Mcpherson Pt. - J M 5 - Lord Bight 222 A M 6 July 13 Hazardous Cove 150 J M 7 July 14 No-Name Pt. - A M 8 July 14 Hazardous Cove 58 J M 9 July 15 Mcpherson Pt. 102 J M 10 July 15 Mcpherson Pt 52 J M 11 July 15 Mcpherson Pt. - J F 12 July 15 No-Name Pt. 390 A F 13 July 16 Mcpherson Pt 120 J F 14 July 16 No-Name Pt. 43 J M 15 July 18 No-Name Pt 351 A F 16 July 18 Hazardous Cove 278 A M 17 July 18 Mcpherson Pt. 373 A F 18 July 18 Mcpherson Pt. 62 J F 19 July 19 Mcpherson Pt. 21 J M 20 July 19 Mcpherson Pt. 20 J M 21 July 20 No-Name Pt. 35 J M 22 July 20 No-Name Pt. 25 J -23 July 20 Henslung Cove 285 A M 24 July 20 Hazardous Cove 220 A M 25 July 20 Hazardous Cove 35 J M 26 July 20 Fury Bay 326 A M 27 July 20 Fury Bay 225 v A F 28 July 20 Fury Bay 26 J -29 July 20 Fury Bay 22 J M 30 July 20 Fury Bay 25 J M 31 July 21 Hazardous Cove - - -32 July 22 Henslung Cove - - -33 July 24 Hazardous Cove - - -34 July 25 Hazardous Cove - - -35 July 25 Hazardous Cove - - -85 CO CD CD .TO I to _CD CD CO CD 32 co - Q M— O CO -*—• c CD E CD L_ Zl CO CO CD E "co o 'en o o o E •a <D -•—< o CD LO "CD 2 CO °2 £ "a co c - co S » •g co co co o _o to" CD _ c CO O a -o CL CM CO -Q 'to ^ CO T3 c CD a o Q. =5 < -32 CD o E -JZ o O co g> E to a> to i! ° 11 O x -2 E roE = a. E m CD F a CD to CD CO < o X CD CO .2> CD CD C o CO o o CD -4—' CO Q LO LO LO T - LO LO co LO "fr CN LO "fr LO T — y- LO co LO > CD d d CM d co d CM d d — ' d T — d d d d d d d LO LO CD CD CD CD CD CD CD LO CD CD CO CD LO LO LO LO LO CD CD o CO co 'fr _^ LO co CD CM _^ CO o LO CD 1 1 • CO 1 d d T — CO d d d d d d d d d d d d d CN CN CO CO CO CN CN CN CN CN CM CM CM CN CN CM CM CM CM o LO CD CD h~ o CO CN "fr h- CD CO LO LO CM o "fr CO T — CD CO co CO LO LO CO LO co CD "fr -si- CD co h- LO CO "fr LO CM LO "fr CD "fr CO CD "fr CD CN CD CM co oo CD LO LO CM LO CD co co o 1 o O I O) 0) CO o o CD CO LO CD CD CM CO O) CO T — CO "fr LO LO "fr LO "fr d d d d "fr d d "fr d d "fr d d d "fr "fr p "fr CN o CN o o 0) co CM 'fr co "fr CM CO CD LO CM co o CM CO O d d d d d d d d d d d d ^—- "fr co "fr co "fr co "fr "fr co "fr "fr "fr CO "fr "fr CO -sl- "fr co "fr "fr "fr CO CN LO CN CM LO CM o o o CD o CD CD CD CD co co o oo LO LO d LO CO CO CM 00 CN d "fr d d CM CM CM "fr* CM d d d d 00 CO co co co CO co co co CO CO CO CO CO CO CO CO CO CO co co CO CO < < SA < < o < > SA < < < < < < SA < SA > > o < > LL "fr CM CJ) "fr -fr d c "o CO CO *zt C *-£ >> o .£ .£ O CO Q_ o o CD = C- CD CD 5 CL CQ CD E co c o CO c c D1±IZ CLx; a> a> co Q ^  111 CLJC SZ O CL CL LO CD N >^  >» >* >, CO CD CO T — t - CM CO LL LL CD "fr LO CN d d "fr d ^—< ^—< -+—< c c c c 'o 'o 'o 'o DL Q_ 0. a. CD CD CD CD E E E E CO CO CO CO z z z 6 o 6 6 z z z z ^_ CO "fr ^—' ^—' -*—< to CO CO to Zl Z l Z l "fr "fr "fr CO "fr "fr CM O T -"fr d "fr c ro o CD £ c o Q_ c >-'o Q_ .55 <DCD -k CD o x: o ro 00 1 b CD E ro z I o z c o 0. to T3 c < c o a. CD O -C o O c 'o Q_ co C < c 'o D_ CO c 'o a. ro 'c CD P cpQ ro J Z 5 Q . CD > o co o c CD T3 CO CD C _Z1 to c CD X z i •='•5 -5 "5 0 ) 0 0 ) 0 ) ^ ^ < < < < 0 O 0 0 T - CN CO "fr LO LO CD CD T — T — v - "- T — T — T — CM CM CM ^—' -•—' +-» ^ — ' -*—' -4—' -+—< -+—» CO to CO to CO CO to to CO CO CO CO Zl Zl Zl Zl Zl Zl Zl Zl Z l Zl Z l Z l CD CD CD CD CD CD CO CD CD CD CD CD Zl Zl Zl Zl Zl Zl Zl Z l ZJ Zl Z l Zl < < < < < < < < < < < < CM c o "fr LO CD co CD 0 CM c o CM CM CM CM CD CO co c: CD O CO ! o O CO 2 o c I— 0 •*—» (O 0 .c •e o o co c 0 E 0 i— zz CO CD 0 E "CD O CO o o XT CL i_ o E T 3 0 •4—» O _ 0 0 (0 -o 1 0 ro <» co . C T3 O C t i ro 8 « 3 E co o § E 2 o ro t-X = E ro o X Q. 2> E -> o O CO 3 Q> I CD T3 Q £ CD co ro t ro | CU c n CD re I— c x o L_ -t—< cu u Q. (D < o o o o •* CN to CD OO co cn CD CN O co CD 00 CM co CN CD t m CO CO LO •* CO CO CO CO "* CO o d O o o O O o o o d O o o d d d d d d d d CN 00 1^-CO co 03 h- CO o LO O CO CO CO co 00 T - 00 •* CD CD CN CO o CO •>* co oo o o LO LO CO CO co CO CN CO CO co LO CD 1^-t~- CD CO io LO LO CD LO CO CD co LO CN LO CN LO oo LO* oo LO o CD CD LO CD CD CD CD CN CO CN l> LL 0 0 0 3 0 0 CD T - 00 CN CO CO O LO CO CN CO CO LO O O 03 CO o CO oo oo CO o o CO CO CO CM CO CO CD CO CD CO CO CO T -co o 5 TJ T3 TJ T> T3 T3 c C C C C C C ra CO co CO CO CO co CO CO co <0 <0 in CO >> >> >. >. >. >. >> o o o o o o o 13 3 3 3 Z3 3 3 _ l _ l _ l _J _ l _J _ l o 0. CD E CO Z • o 2 c o a. 0 E ra Z c c c c c c "o o o o o "o CL a. a. a. a. D. c c c c c c o o o o o o !2 !2 CO a> cu cu cu cu cu SZ sz sz x : x : CL CL CL CL a. O O o o o o 2 2 Z> Z> 2 o O C3) c 3 CO c cu X P T3 T3 r- co ra CN CN CN LO LO CO LO LO co 00 CO LO oo CO CO CD CO CD CN CO a> CO CO CO od LO LO LO LO CO CD LO CN CN CN CM CN CN CN CN CN CM CN CN CN CN CN CN CN CN CU •a co a o 0 o sz o O o O O O CM CN CN CN CN CO T - T — CN CN CN CN CN CM >. >> >> >> >. >. >. >. >, 3 3 3 3 3 3 3 3 3 3 —> —3 —> 3 - 3 —3 —} - 3 —3 —3 ^ ^ 3 3 < 10 3 D) 3 < CO 3 D) 3 < 3 < CO 1^  1^  T — T T — T — CO CO CO CO 3 3 3 3 3 D) D) CD ra ra 3 3 3 3 3 < < < < < E o § E 2 o co 1 si 03 X l» El II I.E -E o sz O CU C/3 cu 0) ra Q 1 CD CM CO d d d 8 ^ CO CD CD CM CD CM LO CO CM CD O O r- x-00 •>+ c 'o CL 0 o SI o o T3 c ra ra m ro * i _ o>, CO LU t— CD T— CN t— m CD if) 3 3 O) O) o Z3 3 3 < < < 00 00 Chapter 5. The Uptake of Brodifacoum by Invertebrates Feeding on Bait Containing the Rodenticide Brodifacoum or Norway Rat carcasses Poisoned with Brodifacoum. 5.1 Introduction Invertebrates play an important role in the nutrient cycl ing process by consuming animal ca rcasses and organic matter, releasing nutrients for primary consumers . The bait used as the rodenticide carrier and attractant provides a rich and readily avai lable food source for invertebrates (Stejskal et a l . 1994). Similarly, animal ca rcasses resulting from brodifacoum poisoning are likely attractive to carrion insects. Invertebrates that feed on the bait and/or carcasses , therefore, may be a secondary or tertiary poisoning risk to non-target spec ies that may otherwise be at low risk of poisoning. To-date, the uptake of brodifacoum residues by invertebrates from anticoagulant poisoned carcasses has not been studied. The objective of this study was to identify the invertebrates that are attracted to and consume brodifacoum poisoned rats and the bait blocks within bait stations. The specif ic hypothesis tested was that invertebrates consuming brodifacoum poisoned rat ca rcasses and/or bait would carry detectable levels of brodifacoum. In 1994, different methods of collecting and preserving invertebrates for brodi facoum assay were tested as there was no establ ished method avai lable in the literature. The results al lowed me to establ ish a protocol for t issue collection. In 1995, invertebrates were again col lected using the protocol establ ished and ana lysed for brodifacoum residue. 89 5.2. Materials and Methods 5.2.1 Invertebrate Collections 5.2.1.1 Collection of Invertebrates at Bait Stations In 1994, invertebrates found in bait stations on Lucy Island were col lected and frozen (- 5 Cels ius) individually in plastic bags. In 1995, on Langara Island, bait station operators were requested to document the presence of any invertebrates present in stations. Samp les of invertebrates found in the bait stations were collected during the baiting campaign. The samples were pooled and frozen for analysis (See below). 5.2.1.2 Collection of Carrion Insects In 1994, different methods of collection and preservation of invertebrates exposed to brodifacoum were tried: 1. Preservation in 95% Ethanol 1994 pilot study: Ethanol is an effective preservative for t issue samp les when they cannot be quick frozen in the field. However, it is a lso a solvent and may leach brodifacoum from tissue samples. Rat ca rcasses were laid out in three habitats: along the shorel ine in Nootka reed grass, in the ancient murrelet colony, and in the interior. Within each location, two carcasses , one snap trapped rat (unpoisoned) and one brodifacoum poisoned rat were spaced approximately 50 m apart, and covered with a cage made with 90 hardware cloth (1 cm x 1 cm mesh) to exclude avian scavengers . Each carcass was checked daily or every other day until sufficient numbers of insects were present feeding on the carcasses . The carcass and a shovel full of soil around the carcass were carefully placed into a (30 cm X 20 cm) wooden box and wrapped with a cotton shirt to allow for gas and moisture exchange. Insects were al lowed to exit the box from a 5 cm diameter hole through a down-pipe elbow into a g lass jar filled with 9 5 % ethanol. Insects were al lowed to consume the carcass for approximately 12 wk before the lid was removed, and remaining insects col lected and placed into the ethanol filled jars. 1995: carrion insects were again col lected using the previous boxes. The des ign of the box was modified to minimise handling of carcass and insects to prevent contamination. Three brodifacoum poisoned rats were p laced individually into boxes with a shovel full of soil and covered with noseum netting. A funnel was placed on one end for an entrance. The exit opening was a 5 cm diameter downpipe elbow into a jar filled with 9 5 % ethanol. Only insects that voluntarily left the carcass into the ethanol were ana lysed. Due to the low number of insects col lected, insects from all three samples were pooled for analysis. 91 2. Preservation by Freezing 1994 pilot study: The Langara Lightstation staff offered the use of a -5 Ce ls ius freezer and I dec ided to run the insect collection trials again and freezing the samples. The fresh collected carrion insect experiment was conducted between July and August 1994. Three brodifacoum poisoned and six unpoisoned rat ca rcasses were each placed in a 10 cm diameter, 30 cm long P V C drainage pipe tubes (Figure 5.1). The tube had a ventilation hole cut in the top and covered with noseum netting, while smal l holes were cut in the bottom to prevent accumulat ion of rainwater. A funnel with the same diameter as the tube was placed over one end for an entrance. On the opposite end, a 2.5 cm diameter c lear P V C tubing (10 cm long) was connected to the tube and led into a g lass jar. Insects that were in the g lass jar were col lected every 7 d for 4 wk, sorted by spec ies, and frozen for future analys is. Due to the low numbers of individuals col lected each day, individuals of each spec ies were pooled across time for brodifacoum residue analys is. 3. Preservation by Hexane Cleaned Aluminum Foil and Freezing The results of the carrion insect collection trials in 1994 revealed that there was contamination of control samples from both ethanol and frozen preservation. In 1995, carrion insects were opportunistically col lected from 92 CO - 3 E co O CO S ( 5 E o o c 2 E o o CO E o o CO C0T3 CO C CO E co o gV- ,_ co c\i .VL™, " 1 rn CO CO cS CO CO O ± 3 _ C 0 J 0 c n CO-n CD CD CZ $ £ <D C D -co ra-C-iz CD ro -I-*- C S-g . EE CO — CO P =J •c6"° a co g CO CD o co 2 -b-fz-c=-5 CD CO _ £ C 0 > cog-2S>8 g cu.ratS a> W-Cl CD co CD 5 •D CDi? O - O M- P 0 CO CO o CD O . r z . b CD CD-S CD 'JLZ CO c C CD - > CD CO > 5 l £ ZJ o — CO C D £ _ Q CD • cz co "g C C D > *- ZJ cz o / > u _ • •-—•- •— , CO O 1*-".^ JZ CD l u g I t — CD CD .55 c CD tz_rz _ t co — 1 0 ^ C D ~ CD - t t ro ^ SlT3 C D _ CZ S i S I § 93 brodifacoum poisoned rat carcasses . Surface feeding carrion insects were col lected with chemical ly c leaned (hexane) instruments and wrapped in chemical ly c leaned aluminum foil to prevent contamination. The packets were sl ipped into pre-labelled whirl-pack bags and frozen. 5.2.2 Sample Preparation Samp le s were shipped frozen to N W R C , Hull, PQ . After the samples were prepared, one-third (by weight) of all t issues were archived at N W R C . 5.2.2.1 Bait Station Invertebrates The snai ls and slugs had their shel ls removed and the soft t issue we ighed. The shel ls were rinsed with methanol in a 1 mm diameter kitchen strainer into a cup holding the soft t issue. 5.2.2.2 Carrion Insects 1. Samples Preserved in 95% Ethanol The insects were separated and the ethanol was al lowed to evaporate from the insects for approximately 1 h and the contents weighed. The insects were homogenised and an aliquot of both insects and the ethanol was ana lysed for brodifacoum residue (See below). 94 2. Frozen Samples The insects were removed from containers, counted and weighed. Methanol was used to rinse any brodifacoum residue from original container. The insects were chopped into fine pieces with chemical ly c leaned sc issors to facilitate homogenisat ion. 3. Samples Frozen in Hexane Cleaned Aluminum Foil The insects were separated from the foil packages and weighed. The foil was rinsed with methanol into a cup holding the insects. Insects were chopped with chemical ly c leaned sc issors into fine p ieces to facilitate homogenisat ion. Homogenisat ion was done in chemical ly c leaned g lass jars. A chemical ly c leaned shaft homogeniser was used for all samples . 5.2.3 Brodifacoum Assay Extraction of brodifacoum residues from prepared samples was carried out at NovaMann International, M iss i ssauga, Ontario, following the procedure establ ished by Hunter (1983). The limit of detection was 0.01 ppm. A s part of the quality control, c lean invertebrate and C a n a d a goose [Branta canadensis) liver samples were fortified with brodifacoum and shipped along with treatment invertebrates in 1994. In 1995, c lean snai ls and s lugs were col lected from G raham Island, 1 km across from Langara Island, and fortified with brodifacoum in the field. Samp les were individually wrapped in chemical ly c leaned foil packets and 95 sl ipped into pre-labelled polyethylene whirl pack bags. These control samp les were handled the same way as those invertebrates col lected from rat ca rcasses and bait stations. Al l original chromatography graphical tracings from the instrumental ana lyses were verified by Bryan Wakeford at the N W R C laboratory. 96 5.3. Results 5.3.1 Brodifacoum Assay Quality Control: In 1994, the brodifacoum residue recovery from fortified samp les was 14% from insect samples and 0.0% from the liver sample (Table 5.1). In 1995, recovery of brodifacoum ranged between 0.0% and 127% (Table 5.2). All control samples did not test positive for brodifacoum. Table 5.1. Brodi facoum residue recovery from fortified samples, Langara Island, 1994. T i ssue Type Fortified Level T issue Brodi facoum % Recovery ug Brodi facoum Weight g Reported ug Liver 5.4 2.2 None Detected 0.0 Insect 9.9 4 1.4 14% 97 Table 5.2. Brodi facoum residue recovery rate from fortified samples prepared on Langara Island, 1995. Sp ike ug Fortified Brodi facoum Percent T i ssue Type Solution Brodi facoum Level Reported Recovery Added (ul) Added (ug/g) (ug) 115 ug/ml Vespericola sp. 56 6.4 0.30 7.16 111.2% Snai l Vespericola sp. 54 4.14 0.34 5.26 127.1% Snai l Vespericola sp. 100 11.5 0.40 11.91 103.5% Snai l Vespericola sp. 100 11.5 1.46 5.42 47 . 11% Snai l Ariolimax sp. 100 11.5 0.85 0.00 0.00% Slug Ariolimax sp. 74 8.51 - a Slug Ariolimax sp. 90 10.35 0.34 3.02 29 .2% Slug Ariolimax sp. 52 5.98 0.26 4.00 66 .9% Slug Sp ike Solut ion 115 ug/ml 93.14 81 .0% Haplotrema sp. 0.00 0.00 0.00 0.00 b Control Snai l Ariolimax sp. 0.00 0.00 0.00 0.00 b Control S lug Ariolimax sp. 0.00 0.00 0.00 0.00 b Control S lug a Not Processed because of leakage of contents in transport, b Control Sample. 98 5.3.2 Bait Station Invertebrates The terrestrial molluscs, banana slugs (Ariolimax sp.) and snai ls (Vespericola sp. and Haplotrema sp.) were the most common and abundant invertebrates found in bait stations and feeding on the bait. The blue coloured bait could be readily seen through the translucent body on those individuals found feeding on bait. B lue casts from both snai ls and slugs were evident around stations on leaves and moss, a s well as inside stations. The mean daily reported proportion of bait stations with s lugs and snai ls were similar at 8.9% and 8.2% respectively (two-tailed t-test; P>0.05) (Figure 5.2). The proportion of bait stations with one or more snai ls ranged from 1.9% to a high of 17.4%. The brown shel led snai l, Vespericola sp. was reported in a mean of 5.8% bait stations daily, significantly more than the yellow/green coloured snai l, Haplotrema sp. (3.0%) (one-tailed t-test; P<0.05) (Figure 5.3). The brodifacoum residue levels for the terrestrial mol luscs ranged from a low of 0.54 ug/g in banana s lugs to a high of 4.13 ug/g in Haplotrema sp. snai ls (Table 5.3). Other invertebrates found in bait stations included crickets (Order Orthoptera) two mil l ipedes (C lass Diplopoda, Harpaphe sp. and one unidentified) and harvestmen (Order Phalagida) although they were uncommon. The blue coloured bait could be seen through the translucent body of the crickets found in bait stations. The cylindrical mil l ipedes burrowed into the bait blocks. Genera l ly the mil l ipedes were found burrowed into the bait blocks alone, however they were also 99 0,18-,-15 16 17 18 19 20 21 22 23 24 25 26 27 28 31 (412) (353) (367) (352) (328) (267) (444) (333) (552) (312) (397) (306) (369) (264) (139) Date, July 1995 Figure 5.2. Proportion of bait stations with snai ls and banana s lugs over the course of the intensive baiting period, Langara Island, 1995 (number of bait stations in brackets). 100 0.12 T 17 18 19 20 21 22 23 24 25 26 27 28 31 (367) (352) (328) (267) (444) (333) (552) (312) (397) (306) (369) (264) (139) Date, July 1995 Figure 5.3. Proportion of bait stations with the terrestrial snai ls, Vespericola sp. and Haplotrema sp., over the course of the intensive baiting period (number of bait stations brackets). 101 Table 5.3. Brodi facoum residues in invertebrates found in bait stations, Langara Island, 1995. Common Name Taxonomy No. Individuals Mean Weight g Brodi facoum ug/g ug Brodi facoum per individual Banana S lug Ariolimax sp. 13 13 10.2 17.1 2.90 0.54 29.55 9.21 Snai l Vespericola sp. 13 40 0.85 0.62 1.53 1.42 1.29 0.88 Snai l Haplotrema sp. 9 30 15 1.16 0.87 0.91 3.57 4.13 4.04 4.12 3.58 3.66 Cricket Order Orthoptera 10 0.04 8.17 0.33 Cyl indrical Mil l ipede C l a s s Diplopoda 30 0.08 2.74 0.22 Mil l ipede Harpaphe sp. 4 0.2 3.70 0.74 Predatory Cent ipede C l a s s Ch i lopoda 1 0.2 0.0 0.0 Ground Beet le Family Carab idae 3 0.1 0.0 0.0 Daddy Order 7 0.014 266.4 3.81 Long Legs Pha lang ida 102 reported in groups of three or more. Harvestmen or daddy long legs were found to contain the highest concentration of brodifacoum (Table 5.3.) Only terrestrial mol luscs and one mill ipede were found to feed on bait in stations on Lucy Island in 1994. Brodi facoum residues could not be detected in Haplotrema sp. snai ls but were detected in Vespericola sp. (Table 5.4). Brodi facoum residues in banana slugs were at detection limits. Table 5.4. Brodi facoum residues in invertebrates found in bait stations, Lucy Island, 1994. Common Name Genus No. Individuals Mean Weight (g) Brodi facoum (ug/g) ug Brodi facoum per individual Snai l Vespericola sp. 46 0.40 2.462 0.984 Snai l Haplotrema sp. 6 0.27 0.000 0.000 Banana S lug Ariolimax sp. 8 3.94 0.002 0.002 Mil l ipede Harpaphe sp. 1 0.56 1.607 0.900 103 5.3.2. Carrion Insects 5.3.2.1 Preservation in 95% Ethanol There were four genus that were the most abundant at rat carcasses; these include three genus from the Order Coleoptera (Beetles): Nicrophorus sp., Necrophilus sp., and Catops sp. The fourth group were the blowflies or blue-bottle flies from the Order Diptera, Family Cal l iphoridae, Genus Calliphora. In 1994, brodifacoum residues were detected in the pooled insects col lected from brodifacoum poisoned and control rat carcasses (Table 5.5). Brodi facoum residues were also detected in the ethanol used for the preservation of insects. Brodi facoum residues were not detected in insects col lected into ethanol in 1995 (Table 5.6). 5.3.2.2 Preservation by Freezing In 1994, brodifacoum residues were detected only in Nicrophorus sp. larvae (Table 5.7). Brodi facoum residues were not detected in the more common and abundant blowfly larva or adult forms. Brodi facoum residues were detected in adult blowflies col lected from a control carcass (Table 5.7). 5.3.2.3 Preservation in Chemically Cleaned Aluminum Foil and Freezing Brodi facoum residues were detected in the blowfly larvae col lected from the radio-collared Norway rat ca rcasses in 1995 (Table 5.8). The concentration was variable and ranged from 0.27-11.39 ug/g larva. Smal l ants (Family Formic idae) were col lected from two rat carcasses, but no brodifacoum residues were detected. 104 l l oo CO CN +-» CO or CO or c o c o o o " t LO T - LO CN C O C N ^ C N O O C N $ 2 o O O O O) CM O " t CN t -" t LO CD CM CD si O O CO o ^ CD CD W CD i i b TO CO CO O — c 2 8 | CO 2 o E £ ZJ - C 8 o .CO v O " D LO O CD -D .F C . CO CO "5 2 co CO CO CD T - o o CN CN CO CD O E T - T - O O O CM CO •si- CD CO CN CM CO O O CO CD CO • O if) o> 8 •S ~° (0 c h- CO E o c o X CO CD • 3 C O CO Zl c CD o ±i CO ±t ±^ CO ±i CO ±i Z J ^ Z l Z l ^ Z i ^ Z l 5 c 6 5 5 c o 5 c o 5 CD T3 C L CO CO ! CD CO i— £ Q . o _CD O O C L CO CO c L 5 CO O co -§ co ^ CL co 2 o •c .Q. "co O CO 1_ CD Q . b " O CD c CD -g 'c Z) CO o 3 D) 'CD CO •4-* o CD CO c E Zl o o 2 CO o c CO L U E Zl o o O 1_ CD Table 5.6. Insects col lected and preserved in 95% ethanol with no brodifacoum residues detected, Langara Island, 1995. Order Family/Genus Common Name Association Number Coleoptera Catops sp. Carrion beetle Carrion 184 (Beetles) Necrophilus sp. Carrion beetle Carrion 2 Nicrophorus sp. Carrion beetle Carrion 1 Family Rhizophagidae Root-Eating Decaying 3 beetles wood Family Dermistidae Skin beetles Carrion & 18 Vegetation Unidentified - - 54 Diptera Calliphora sp. Blowflies Carrion 4 (Flies) Family Ptychopteridae Phantom Crane Decaying 5 flies Vegetation Family Trichoceridae Winter Crane Decaying 1 flies Vegetation Family Phoridae Humpbacked Decaying 7 flies Vegetation Family Platypezidae Fiat-Footed flies Vegetation 3 Family Asilidae Robber flies Insect 1 Larva Family Anthomyiidae Anthomyiid flies 2 Unidentified - - 2 Collembola Springtails Soil >63 Diplura Diplurans Soil 4 Lepidoptera Moths Vegetation 1 Araneae Spiders Insects 5 Acari Spiders Insects 2 Total Brodifacoum (mg/kg) >362 None Detected 106 Table 5.7. Brodi facoum residues in carrion insects col lected from brodifacoum poisoned Norway rat carcasses , col lected fresh, Langara Island, 1994. Genu s Life Mean Brodi facoum ug Stage No. Weight (ug/g) Brodi facoum (g) per Individual Nicrophorus sp. Larva 13 Necrophilus sp. Adult 18 1.02 0.18 0.860 N D a 0.877 < Detection Limits Catops sp. Adult 193 0.002 ND < Detection Limits Calliphora sp. Larva 204 0.02 ND < Detection Limits Calliphora sp. Adult 25 0.14 ND < Detection Limits a None Detected Table 5.8. Brodi facoum residues in blowfly larva col lected from brodifacoum poisoned Norway rat carcasses, Langara Island, 1995. No. Larva Mean Larva ug Brodi facoum Brodi facoum in Pool Weight (g) per Individual Concentrat ion / Larva ug/g Larva 100 0.05 0.24 4.81 100 0.06 0.15 2.65 15 0.05 0.12 2.54 56 0.06 0.02 0.27 40 0.02 0.26 11.39 25 0.09 0.18 2.09 33 0.08 0.06 0.75 60 0.02 0.05 2.28 107 5.4. Discussion In 1995, collection and preservation protocols were establ ished to rectify the contamination problem encountered in 1994. However, the level of. brodi facoum residue detected from pooled samples was highly variable mostly because of the variable recovery rates from sample to sample by the as say procedure used. The procedure for assay ing the 4-hydroxycoumarin anticoagulants has been publ ished with mean recovery rates between 89-91% (Hunter 1983). However, the results provided by NovaMann Laborator ies showed recovery rates ranging from 0 to 127% from the spiked samples. No explanation was offered by NovaMann. Another factor which may have contributed to the variability is that the brodifacoum residues detected may represent unassimi lated brodifacoum in the gut of some invertebrate samples and assimilated brodifacoum into the t issues of other samples . The control samples did not test positive for brodifacoum, thus, I am confident that the qualitative brodifacoum residue analysis was sound and brodi facoum was detected from samples in 1995. The data presented indicate that invertebrates feeding on bait and brodifacoum poisoned rats did contain detectable levels of brodifacoum residue. Carr ion insects were a potential tertiary poisoning risk and the invertebrates feeding on the bait were a ' secondary poisoning risk to non-target spec ies. In New Zea land, invertebrates were observed to feed on baits, and brodifacoum residues were found in beetles 1 0 8 col lected from bait stations containing bait intended for rats on Stewart Island, however, no data were presented (Eason and Spurr 1995). The only reported secondary poisoning of insectivorous birds was in a zoo where birds in an aviary died after feeding on ants and cockroaches that had eaten bait containing brodi facoum (Godfrey 1985). The blue coloured bait could be readily seen through the translucent bodies of the snai ls and slugs found feeding on the bait. The brodifacoum res idues in the gut of Norway rats represented between 30-50% of the whole body residue level (See Chapter 4). The blowfly larva col lected likely had a gut full of carrion containing brodifacoum residues. The implication is that the secondary/tertiary poisoning hazard to non-target spec ies is greater from an invertebrate that has been recently feeding on bait/carrion containing brodifacoum versus one that has had time to excrete the contents containing brodifacoum. In other words, the poisoning risk can be defined as short and long term. The short term risk may be greater than the long term risk because of the presence of brodifacoum in the gut of the invertebrate. In order to elucidate the above possibility, known amounts of brodifacoum could be fed to invertebrates and the residue levels from pooled or individual samples can be ana lysed at var ious time points after feeding including when the bait has passed through the gut. This would allow for measurement of the brodifacoum res idues in the gut versus levels of brodifacoum retained in the t issues. 109 The attractiveness of the bait to snai ls and slugs indicate that any excess bait d ispersed by rats over the course of the intensive baiting period would be consumed. The amount of bait d ispersed by rats and not consumed is unknown. However, bait crumbs were found in and around burrows, under logs, and along runs used by rats. On Lucy Island in 1994, a Vespericola sp. snail was found to be feeding on the bait crumbs outside a burrow and had excreted a blue cast. In May 1995, 9 months after the removal of the bait from the stations, four old bait blocks (9.6 g) were found under a log on the west side of Lucy Island and contained 10.986 ug/g brodifacoum. Snai ls found near the bait were collected and contained a concentration of 0.910 mg/kg brodifacoum. Thus, the snai ls and s lugs could be an ongoing secondary poisoning risk until the bait has degraded or is consumed. No brodifacoum residues were detected in the ethanol preserved insects in 1995. This may be due to the low consumption of the ca rcasses by carrion insects. The ca rcasses were not being consumed by carrion insects as found the previous year using a similar des ign trap. This may be due to the high number of other wild rat carcasses that were avai lable to insects or the altered des ign of the trap attracted fewer carrion insects. Most of the other insects were assoc iated with vegetation and soil, and were a result of placing a shovel full of soil in with the carcass . The brodifacoum residue data from blowfly larva col lected in 1995 were more reliable due to the different collection technique and increased attention to 110 minimising or eliminating contamination due to collection. It can be conc luded that blowfly larva consuming rat carcasses containing brodifacoum pose a tertiary poisoning risk to non-target spec ies. In 1994, song sparrows and Northwestern crows were photographed at rat ca rcasses that had been attacked by carrion beetles (Chapter 4). They were at risk of tertiary poisoning in 1995, if the carrion insects consuming ca rcasses did carry a load of brodifacoum residue. Poo led samples of song sparrow livers tested positive for brodi facoum (Appendix 2). Unfortunately, the highly variable quantitative brodifacoum residue data preclude estimating actual risk of poisoning from the invertebrates. The confirmation of brodifacoum residues in invertebrates indicates that insect ivorous non-target spec ies may be exposed to brodifacoum and need to be cons idered in future eradications. The saturation baiting strategy used on Langara Island provided a constant food supply for snai ls, s lugs and other invertebrates. The carrion insects rapidly and readily consumed brodifacoum poisoned rat ca rcasses and subsequent ly attracted song sparrows and Northwestern crows (Chapter 4). The above data indicate that to minimise potential non-target spec ies poisoning, less brodifacoum in the form of bait and/or po isoned carcasses should be avai lable to invertebrates. In conclus ion, brodifacoum residues were detected in all invertebrates attracted to bait in the stations. The positive detection of brodifacoum residues in carrion insects and bait station invertebrates indicates that they were a tertiary and secondary poisoning risk to insectivorous spec ies. 111 Chapter 6. Conclusions and Recommendations The overall objective of this thesis was to investigate the short term poisoning hazard to non-target spec ies from brodifacoum bait used to eradicate rats from Langara and surrounding Lucy and Cox Islands. Brodi facoum residues were detected in every level investigated: carrion insects, terrestrial mol luscs, songbirds, ravens, Northwestern crows, and bald eagles - the top of the food chain. The results presented in this thesis provide the basis for future work by identifying pathways and spec ies at risk of poisoning during similar operat ions along the British Co lumbia coast. 6.1 Conclusions 6.1.1 Native Small Mammal Study 1. Whi le Dusky shrews were attracted to the bait in the stations, the decl ine in their population was non-significant in all regions on Langara Island, indicating they were at low risk of extirpation. 2. Adult breeding Dusky shrews appeared to be at greater risk of poisoning than non-breeders and juveni les. 6.1.2 Secondary Poisoning Risk to Avian Scavengers and Predators 1. Rats dying above ground are a consequence of anticoagulant poisoning and is not age or sex related. 112 2. The mean whole body brodifacoum residue concentrations in Norway rats found dead above ground were the highest yet reported in the literature. 3. C o m m o n ravens were at an extreme risk of primary and secondary poisoning: i. Ravens were able to gain access to the bait in the stations. ii. Ravens were identified as the most significant scavenger of Norway rat carcasses . 4. Ba ld eag les were exposed to brodifacoum, however, no mortality was detected. 5. Brodi facoum was detected in Northwestern crows 9 months after the Lucy Island baiting in 1994. 6.1.3 Study of Brodifacoum Transport into the Ecosystem 1. The bait blocks were highly attractive to terrestrial snai ls and s lugs. 2. Brodi facoum residues were detected in snails, s lugs, blowfly larva, and other spec ies, however, it is unclear if it was unassimilated brodifacoum in the gut and/or brodifacoum retained in the t issues. 6.1.4 Detection of Brodifacoum Exposure through Plasma Residue Analysis and Prothrombin Time Evaluation. 1. P l a sma brodifacoum residue analysis is an effective indicator of exposure in birds. 2. The use of a mammal ian derived thromboplastin is only an effective tool to detect a severely affected bird. 113 6.2 Recommendations The following eight recommendat ions are presented to minimise the non-target spec ies poisoning during similar rat eradication operations in the future. They are presented from a non-target spec ies perspective and some recommendat ions may not be economical ly or operationally feasible under all condit ions. 1. Switch from saturation baiting to pulsed baiting. Utilising a pulsed baiting strategy, where the bait is avai lable in limited quantities and is only replenished at pre-specif ied intervals, has been shown to reduce the residue load of target rodents (Merson et al. 1984; Kauke inen 1982). Alternatively, reduction of the concentration of the active ingredient has also been shown to reduce the target spec ies residue level (Kaukeinen 1982). The pulsed baiting strategy would a lso reduce the total amount of brodifacoum released into the environment, and avai lable for potential transport into the ecosystem by invertebrates. The cost is an increased time to eradication which could translates into an increased economic cost. 2. Shift the intensive baiting period to the time of the year when the rat population is at its lowest. This would reduce the absolute number of rats dying above ground and therefore, decreas ing availability to scavengers and predators. The late winter months are likely the time rat populations are lowest. However, it may not 114 be feasible to have people working on and around offshore is lands during winter months off the coast of British Co lumbia when severe storms are common. 3. The use of anticoagulants that are less toxic to non-target spec ies should be investigated and considered. Brodi facoum is very effective for controlling rodents due to its high toxicity to the target spec ies. However, it is a lso highly toxic to non-target bird spec ies at dose levels that were readily avai lable during the Langara Island Seabi rd Habitat Restoration Project. The use of a less toxic anticoagulant would reduce the primary and secondary poisoning risk to birds. But it is generally avoided because of the risk of rats developing bait shyness or resistance. On the other hand, there are other anticoagulants that are as toxic as brodifacoum to rats, but less toxic to birds. For example, f locoumafen has a L D 5 0 of 0.25-0.56 mg/kg for Norway rats (similar to brodifacoum) (Huckle et al. 1989b) but is approximately 26 times less toxic to Japanese quail (flocoumafen LD 5 0 = >300 mg/kg (Huckle et al. 1989b); brodifacoum LD 5 0 = 11.6 mg/kg (Ross et al. 1976)). 4. Re-design bait stations to exclude Common Ravens. Studies with different bait station designs such as an S-shape or simply, a longer station should be undertaken to determine the shape that does not compromise acceptance by rats but excludes ravens. 115 5. Use bait formulations that are less attractive to birds. The use of a bait formulation that would make the baits unpalatable to birds but maintains high attractiveness to the target spec ies would be ideal. In the United States, methyl anthranilate is used as an effective non-lethal bird repellent. It has been used effectively as a non-lethal bird repellent in the field (Askham 1994; Avery 1992; Cummings et al. 1991; G lahn e t a l . 1989). It is not yet registered for use in Canada (P. Mineau, pers. comm.), however, further study into the effectiveness of this product, or others, to prevent primary poisoning to birds otherwise attracted to the bait is warranted. 6. Use non-removable bait blocks in stations. Even if recommendation 4 is followed, rats may disperse large quantities of bait and potentially make them access ib le to non-target species. The bait blocks should be fastened down so that they cannot be removed but must be consumed within the stations thus minimising dispersal. Alternatively, the bait could be reformulated into a brick (15 cm x 15 cm x 2.5 cm) that cannot be physically removed from the station but requires rats to feed directly on the bait in station. Either option would reduce the number of visits to a station by an operator, thus, reducing the number of crew required to check stations. On the down side, fixing the bait may lead to dominant rats defending stations and preventing conspecif ics from gaining access to bait. A s a consequence, the time to eradication may be lengthened. 1 1 6 7. Monitor non-target spec ies before, during, and after eradication campaigns. i. Avian scavengers. Common ravens are the non-target spec ies at greatest risk of poisoning. Ravens are common scavengers in seabird colonies in the Queen Charlotte archipelago (Rodway et al. 1990;1988). Their aggressive and inquisitive behaviour put them in the highest risk of poisoning of all spec ies studied and thus, could be used as an indicator of non-target spec ies poisoning. They could indicate the availability of brodifacoum to non-target species. Population estimates before, during and post eradication should be made and compared against one or more control site(s). If an effective trapping method can be developed, blood sampling of adult and juvenile ravens during and post eradication can be used to monitor the exposure of individuals to brodifacoum. If future eradications are pursued in late winter/early spring when ravens are nesting, pre-fledging raven chicks may be a readily access ib le source of p lasma. The plasma brodifacoum residue analysis is an effective, inexpensive indicator of exposure. A laboratory that is already set up for the analysis should be used to ensure quality control. ii. Native small mammals. Many seabird colonies in the Queen Charlotte archipelago harbor populations of native deer mice and/or dusky shrews. Shrews are unlikely to be extirpated from islands following the regime of the Langara Island Seabird Habitat Restoration Project. However, the operation has the potential to alter population dynamics. Smal l mammal populations should be monitored for changes in the short term and long term. Variables such as population size, 117 demographics including age and sex structure changes should be measured and quantified. A control site, with similar small mammal composition and island size, should be monitored concurrently. 8. Training in the safe handling and disposal of bait blocks is essential. At least three of the 13 ravens found dead were primarily poisoned as a direct result of a bait spill. This incident indicates that a strict protocol in the handling and disposal of bait blocks is required. A training program specifically designed to provide training for island restoration crews dispensing and disposing of bait should be developed. Adherence to all or some of these recommendat ions would reduce the non-target spec ies poisoning risk substantially. In particular, if only the first four recommendat ions are fol lowed, the non-target spec ies poisoning risk should already be lessened significantly. 6.3 Future Directions This thesis only investigated the short term impacts of the poisoning operation and does not address the long term impacts. Over 14, 000 bait b locks were d ispensed from bait stations across the island over the intensive baiting period. However, the 1,100 active bait stations were concentrated primarily within a few hundred meters of the shoreline, leaving the possibility of ongoing primary and secondary poisoning risks to non-target spec ies. The long-term effects of the 118 baiting on both the shrew and raven population should be investigated. Endpoints such as mortality should be continued to be monitored but a lso sub-lethal effects such as reproduction. The use of radio-telemetry would be useful to monitor the hazard to individual common ravens. This technology al lows for the determination of the location, time, and potentially, cause of death. If brodifacoum poisoned ravens were preyed on or s cavenged by eagles, radio telemetry could possibly answer this quest ion. Laboratory investigation into the L D 5 0 of brodifacoum to local spec ies, such as the common raven, would be useful for determining risk of poisoning from brodifacoum residues found within spec ies investigated in this thesis. Similarly, laboratory investigation into the biological signif icance of brodifacoum p lasma res idues on prothrombin time (a measure of risk of hemorrhaging) in birds would be useful to interpret residue data. Further investigations into modell ing accumulat ion and transfer of residues, such as that deve loped by Smith et al. (1990), as it appl ies to island restoration should be undertaken so as to increase the availability of the tools in assess ing the non-target spec ies poisoning cost. 119 References Adorjan, A.S. , Kolenosky, G.B. 1969. A manual for the identification of hairs of se lected Ontario mammals . Research Branch. Department of Lands and Forests. A ldenberg, T. and Slob, W. 1993. Conf idence limits for hazardous concentrat ions based on logistically distributed N O E C toxicity data. Ecotoxicology and Environmental Safety. 25:48-63. Anonymous . 1976. Warfarin and fetal abnormality. Lancet. 1:1127. A s kham, L.R. 1994. Eff icacy of methyl anthranilate as a bird repellent on cherries, blueberries and grapes. Proc. Vert. Pest Conf. 15:137-141. Avery, M.L. 1992. Evaluation of methyl anthranilate as a bird repellent in fruit crops. Proc. Vert. Pes t Conf. 15:130-133. Ba lcomb, R. 1986. Songbird ca rcasses d isappear rapidly from agricultural f ields. The Auk. 103:817-820. Bellevil le, J., Corni l lon, J.P., Baguet, J., C lendinnen, G . and Eloy, R. 1982. Hemostas is , blood coagulation, and fibrinolysis in the J apanese quai l . Comp . B iochem. Physio l . 71 A:219-230. Bertram, D.F. 1989. The status of Ancient Murrelets breeding on Langara Island, British Co lumbia, in 1988. Canad ian Wildlife Serv ice, Technica l Report Ser ies No. 59. Pacif ic and Yukon Reg ion. 67pp. Bertram, D.F. and Nagorsen, D.W. 1995. Introduced Rats, Rattus spp., on the Queen Charlotte Islands: Implications for Seabi rd Conservat ion. Canad ian Field Naturalist. 109:(1):6-10. Blus, L.J., Henny, C.J . and Grove, R.A. 1987. Effects of pelletized anticoagulant rodenticides on California quail. Journal of Wildlife D i seases . 21(4):391-395. Bond, C.F. and Gilbert, P.W. 1958. Comparat ive study of blood vo lume in representative aquatic and non-aquatic birds. A m . J . Phys io l . 194:519. Bortolotti, G.R. 1984. Sexua l s ize dimorphism and age-related s ize variation in bald eag les. J . Wildl. Manage. 48(1):72-81. Bowman, T.D., Schempf, P.F. and Bernatowicz, J.A. 1995. Bald eagle survival and population dynamics in A l aska after the Exxon Valdez oil spil l. J . Wi ldl . Manage . 59(2):317-324. 120 Brandt, J.T. 1991. Hemostas is . In: Laboratory Medicine: Test Select ion and Interpretation. By J .H. Howanitz and P.J. Howanitz. Sect ion editors P. J . Cornbleet, R.B. Schif fman, and L.D. Petz. Churchi l l Livingston Inc. Brenchley, A. 1985. Ravens and their effect on sheep production on Saltspr ing Island, British Columbia: A preliminary study. Crop Protection Branch, Ministry of Agriculture and Food, Victoria, B.C. Canada . 27 pp. Brown, B.A. 1988. Hematology: Principles and Procedures. 5th Edit ion. Lea and Febiger, Phi ladelphia. Brown, R.A., Hardy, A.R., Greig-Smith, P.W. and Edwards, P.J. 1988. A s se s s i ng the impact of rodenticides on the environment. E P P O Bulletin 18:283-292. Buck, P.E.F. 1995. Field test of Ratak+ weather block on Lucy Island, British Co lumbia , Canada . Prepared for Environment Canada , Canad ian Wildlife Serv ice, Pacif ic and Yukon Region. Unpubl ished report. 14 pp. Ca in , S.L. and Hodges, J.I. 1989. A floating-fish snare for capturing bald eag les. Journal of Raptor Research . 23(1): 10-13. Caldwel l , P.T., Ren , P. and Bel l , R.G. 1974. Warfarin and metabol ism of vitamin K 1 . B iochemica l Pharmacology. 23:3353-3362. Ca lhoun, J .B. 1962. The Eco logy and Soc io logy of the Norway Rat. U.S. Department of Health, Educat ion, and Welfare. Publ ic Health Serv ice. 288 pp. Campbe l l , R.W., Dawe, N.K., Cowan , I. McT., Cooper, J.M., Kaiser.G.W., and McNal l , M.C.E.. 1990. Birds of British Co lumbia Vol .2. Roya l British Co lumbia Museum.Victor ia, B.C. Choonara , I.A., Mal ia, R.G., Haynes, B.P., Hay, C.R., Cholerton, S., Breckenr idge, A.M. , Preston, F.E., and Park, B.K. 1988. The relationship between inhibition of vitamin K1 2,3-epoxide reductase and reduction of clotting factor activity with Warfar in. Br. J . Cl in. Pharm 25:1-7. Churchf ie ld, S. 1990. The Natural History of Shrews. Comstock Publ ishing Assoc ia tes , New York. 178 pp. Colv in, B.A. and Hegdal , P.L. 1987. Procedures for assess ing secondary poisoning hazards of rodenticides to owls. In: Vertebrate Pest Control and Management Materials: 5th Vo lume. A S T M STP974 . S.A. Shumake and R.W. Bul lard, Editors. Amer i can Society for Testing and Materials, Phi ladelphia. 64-71. 121 Colv in, B.A., Hegdal , P.L. and Jackson , W.B. 1988. Rev iew of non-target hazards assoc iated with rodenticide use in the USA . E P P O Bulletin 18:301-308. Cox, P.R. 1990. Environmental effects of rodenticide use. P h D thesis. University of Read ing. Cox, P. and Smith, R.H. 1990. Rodentic ide ecotoxicology: assess ing non-target population effects. Functional Ecology. 4:315-320. Cox, P. and Smith, R.H. 1992. Rodentic ide ecotoxicology: Pre-lethal effects of anticoagulants on rat behaviour. Proc. Vert. Pest Conf. 15:165-170. Cra ig, V. J . 1995. Relat ionships between shrews (Sorex spp.) and downed wood in the Vancouve r watersheds, B.C. M.Sc. Thes is . Faculty of Forestry. University of British Co lumbia . 88 pp. Cummings , J.L., Mason , J.R., Otis, D.L. and Heisterberg, J.F. 1991. Evaluat ion of dimethyl and methyl anthranilate as Canada goose repellent on grass. Wildlife Soc iety Bulletin. 19:184-190. Didisheim, P., Hattori, K. and Lewis, J .H. 1959. Hematologic and coagulat ion studies in var ious animal spec ies. J . Lab. and Cl in. Med . 53(6):866-875. Doerr, J.A. and Hamilton, P.B. 1981. New evidence for intrinsic blood coagulat ion in ch ickens. Poultry Sc ience 60:237-242. Dorn, P. and Muller, F. Untersuchungen ube rd i e thromboplast inzeit-bestimmung mit homologen und heterologen Thrombokinasen am normalen, Vitamin K-arm ernahrten und Sulfonamid-belasten Huhn. Zentbl. Ve tMed . 12A:380-387. Drent, R.H., and Guiguet, C.J . 1961. A catalog of British Co lumbia seabird Co lon ies. Occas iona l Paper No. 12. British Co lumbia Provincial Museum, Victoria. Dubock, A .C . 1984. Pu lsed baiting- a new technique for high potency, s low acting rodenticides. In: Proceedings of Vertebrate Pest Control. A .C . Dubock, Editor. ICI, Fernhurst, U.K. 105-142. Dubock, A .C . and Kauke inen, D.E. 1978. Brodi facoum (Talon Rodentic ide), a Novel Concept . Proc. Vert. Pest Conf. 8:127-137. Eason , C T . and Spurr, E.B. 1995. Rev iew of the toxicity and impacts of brodifacoum on non-target wildlife in New Zea land. New Zea land Journal of Zoology. 22:371-379. 122 Eason , C.T., Wright, G.R. and Batcheler, D. 1996. Anticoagulant effects and the pers istence of brodifacoum in possums (Trichosurus vulpecula). New Zea land Journal of Agricultural Research . 39:397-400. Ehrl ich, P.R., Dobkin, D.S., and Wheye, D. 1988. The birder's handbook. A field guide to the natural history of North Amer ican birds. S imon and Schuster. 785 pp. Evans , J . and Ward , A.L. 1967. Secondary poisoning assoc iated with anticoagulant-kil led nutria. J A V M A 151(7):856-861. Fal la, R.A., S ibson, R.B. and Turbott, E.G. 1983. The new guide to the birds of New Zea land. Coll ins, Auckland. 247 pp. Fe insod, F.M., A l lam, I., Raga i , A. and Saah , A . J . 1986. Rodent ic ide- induced s igns simulating Rift Va l ley fever in sheep and goats in Egypt. Veter inary Record . 118:270-272. Fenn, M.G.P., Tew, T.E. and Macdonald, D.W. 1987. Rat movements and control on an Oxfordshire farm. J . Zool . London. 213:745-749. Fiore, C.E., Tamburino, C , Foti, R. and Grimaldi, D. 1990. Reduced bone mineral content in patients taking and oral anticoagulant. South Med J . 83:538-542. Foster, J .B. 1965. The evolution of the mammals of the Queen Charlotte Islands, British Co lumbia . B C Provincial Museum Occas iona l Papers No. 14. 96 pp. Gemmeke , von H. 1990. Untersuchungen zur Absch tzung des sekundarvergiftungsriskos bei Beuregreifen durch Rodentiz ic ide. Nachtrichtenb. Deut. Pf lanzenschutzd 42:22-25. Gerrard, J .M. and Bortolotti, G.R. 1988. The Bald Eagle. Haunts and Habits of a Wi lderness Monarch. Smithsonian Institution Press . 177 pp. G lahn , J.F., Mason J.R. and Woods , D.R. 1989. Dimethyl anthranilate as a bird repellent in l ivestock feed. Wildlife Society Bulletin. 17:313-320. Godfrey, M.E.R. 1985. Non-target and secondary poisoning hazards of "second-generat ion" anticoagulants. Ac ta Zoo l . Fenn ica. 173:209-212. Godfrey, M.E.R. 1986. An evaluation of the acute oral toxicity of brodifacoum to birds. Proc. Vert. Pest Conf. 12:78-81. 123 Graham, D. 1977. A f i e l d evaluation of the cyclone seeder for reducing R ichardson ground squirrel populations causing damage in central Montana. Montana Dept. L ivestock Vertebrate Pest Conf. Bureau, Brands Enforcement Divis ion. He lena, Montana. 15 pages. Gray, A., Eadsforth, C.V. and Dutton, A . J . 1994. Non-invasive method for monitoring the exposure of barn owls to second-generat ion rodenticides. Pest ic ide Sc ience 41:339-343. Griminger, P. 1986. Hemostas is . In: Sturkie, P.D. (ed). Av ian Physio logy. New York, NY: Springer-Verlag. 121-129. Hadler, M.R. and Shadbolt, R.S. 1975. Novel 4-hydroxy-coumarin anticoagulants active against resistant rats. Nature 253:275-277. Hall, B.E. and Priestley, I. 1992. Brodifacoum: Metabol ism in soil under aerobic condit ions (IRI Project No. 381441). Haselmere, Surrey, Zene ca Agrochemica ls , 63 pp (Report No. 8795). (Cited in I PCS 1995). Hall, B.E., Jackson R. and Priestley, I. 1992. Di fenacoum: Photolys is in buffered aqueous solutions (IRI Project No. 381614). Hase lmere, Surrey, Zene ca Agrochemica ls , 123 pp (Report No. 8704). (Cited in I PCS 1995). Harfenist, A. 1993. The Status of Ancient Murrelets and other fauna in relation to introduced rats on Langara Island, 1993. Canad ian Wildlife Serv ice, Technica l Report Ser ies No. 218. Pacif ic and Yukon Reg ion. 52 pp. Harrison, E.G., Porter, A . J . and Forbes, S. 1988. Development of methods to a s s e s s the hazards of a rodenticide to non-target vertebrates. British Crop Protection Conference: Environmental effects of Pest ic ides No. 40:89-95. Hawes, M.L. 1977. Home range, territoriality, and ecological separat ion in sympatr ic shrews, Sorex vagrans and Sorex obscurus. Journal of Mammalogy. 58(3):354-367. Hawes, M.L. 1976. Odor as a possible isolating mechan ism in sympatr ic spec ies of shrews (Sorex vagrans and Sorex monticolus). Journal of Mammalogy. 58:354-367. Hegdal , P.L. and Blaskiewicz, R.W. 1984. Evaluation of the potential hazard to Barn Owls of Talon used to control rats and house mice. Environmental Toxicology and Chemistry. 3:167-169. 124 Hegdal , P.L. and Colv in, B.A. 1988. Potential hazard to eastern screech owls and other raptors of brodifacoum bait used for vole control in orchards. Environmental Toxicology and Chemistry. 7:245-260. Hill, E.F. 1994. Acute and subacute toxicology in evaluation of pesticide hazard to avian wildlife. In: Wildlife Toxicology and Populat ion Model ing. R.J. Kendal l and Lacher, J., Editors. Setae Publication, Lewis Publ ishers. 576 pp. Hill, R.W., Maddock, B.G., Hart, B. and Bowles, P.F. 1976. Determination of the acute toxicity of P P 581 to Ra inbow trout (Salmo gairdneri). Br ixham, Devon, Imperial Chemica l Industries Ltd. Br ixham Laboratory (Report No. BL/B/1758). (Cited in I PCS 1995). Hodge, M.C.E., Banham P.B., Richards, D., Weight, T .M. and Wi lson, J . 1980a. Product registration. Brodifacoum: Teratogenicity study in the rat. Geneva Switzer land, Zeneca Agrochemica ls , Stauffer Chemica l B.V. (ICI C T L Report No. CTL/P/437). (Cited in I PCS 1995). Hodge, M.C.E., Banham P.B., Richards, D. and Weight, T .M. 1980b. Product registration. Brodifacoum: Teratogenicity study in the rabbit. Geneva Switzer land, Zeneca Agrochemica ls , Stauffer Chemica l B.V. (ICI C T L Report No. CTL/P/459). (Cited in I PCS 1995). Hodge, M.C.E., Banham P.B., Richards, D. and Weight, T .M. 1980c. Product registration. Brodifacoum: Teratogenicity study in the rabbit (Appendices A-E). Geneva Switzer land, Zeneca Agrochemica ls , Stauffer Chemica l B.V. (ICI C T L Report No. CTL/P/459S). (Cited in I PCS 1995). Hoffman, R.S., Smilkstein, M.J. and Goldfrank, L.R. 1988. Evaluat ion of coagulat ion factor abnormalit ies in long-acting anticoagulant overdose. Cl in ical Toxicology. 26:233-248. Howald, G.R. 1995. Evaluation of primary and secondary non-target spec ies poisoning risk in the Langara Island rat eradication:Prel iminary Report. Prepared for Environment Canada , Canad ian Wildlife Serv ice, Paci f ic and Yukon Region. 23 pp. Huckle, K.R., Morr ison, B.J. and Warburton, P.A. 1989a. The percutaneous fate of the rodenticide f locoumafen in the rat: role of non-biliary intestinal excret ion. Xenobiot ica 19(1):63-74. Huckle, K.R., Warburton, P.A., Forbes, S., and Logan, C.J . 1989b. Studies on the fate of f locoumafen in the Japanese quail (Coturnix coturnix japonica). Xenobiot ica 19(1 ):51-62. 125 Hunter, K. 1983. Determination of coumarin anticoagulant rodenticide res idues in animal t issue by high-performance liquid chromatography. I. F luorescence detection using post-column techniques. Journal of Chromatography. 270:267-276. ICI. 1984. Summary of additional data to support the registration of brodifacoum for rabbits in New Zea land. Imperial Chemica l Industries P L C , Plant Protection Division, Haslmere, England. 21pp. Unpubl ished report. (Cited in Taylor 1993). I PCS . 1995. International Programme on Chemica l Safety. Environmental Health Criteria 175. Anticoagulant Rodent ic ides. World Health Organizat ion, Geneva . 121pp. Jackman , R.E., Hunt, W.G., Driscoll, D.E., and Jenkins, J .M. 1993. A modified floating-fish snare for capture of inland bald eagles. North Amer i can Bird Bander. 18(3):98-101. Jackson , R., Priestley, I. and Hall, B.E. 1991. The determination of the hydrolytic stability of [ 1 4C]-brodifacoum (IRI Project No. 381420). Hase lmere, Surrey, Z ene c a Agrochemica ls . 77pp (Report No. 8330). (Cited in I PCS 1995). Jesty, J . and Nemerson, Y. 1995. The pathways of blood coagulat ion. Ch . 122. pp 1227-1238. In: Wil l iams Hematology. 5th edition. E. Beutler, M.A. L ichtman, B.S. Co l l e r and T.J . Kipps, Editors. McGraw-Hi l l Inc. Health Profess ions Divisions. 1804 pp. Kaiser, G.W., Taylor, R.H., Buck, P.D., Elliott, J.E., Howald, G.R. and Drever, M.C. 1997. The Langara Island Seabird Habitat Recovery Project: 1993-1997. Environment Canada , Pacif ic Wildlife Research Centre. In press. Kase , F. 1978. The effect of Homo- and heterologous thromboplast ins on p lasmas of man, seven mammal ian and two avian species: a comparat ive study. Comp . B iochem. Physio l . 61A:65-68. Kase , F., Butchers, J . and Spurl ing, N.W. 1980. Compar i son of the rate-limiting influence of factor VII on mammal ian p lasma coagulation following extrinsic activation by avian and mammal ian t issue thromboplastins. Comp . B iochem. Phys io l . 65A:421-426. Kauke inen, D.E. 1982. A review of the secondary poisoning hazard potential to wildlife from the use of anticoagulant rodenticides. Proc. Vert. Pes t Conf. 10:151-158. 126 Kie, J .G., Baldwin, J.A. and Evans, C.J . 1994. C A L H O M E . Home range analys is program. U.S. Forest Service, Pacif ic Southwest Research Station, Fresno, CA . Knapen, M.H.J., J ie, K-S.G., Hamulyak, K. and Vermeer C. 1993. Vitamin K induced changes in markers for osteoblast activity and urinary ca lc ium loss. Calcif. T i ssue Int. 53:81-85. Knight, R.L., Randolph, P.J., A l len, G.T., Young, L S . and Wigen, R.J. 1990. Diets of nesting bald eagles, Haliaeetus leucocephalus, in Western Wash ington. Canad ian Field Naturalist 104(4):545-551. Krebs, C . J . 1989. Ecological Methodology. University of British Co lumbia . Harper and Row, New York. Kuehl , R.O. 1994. Statistical Principles of Research Des ign and Ana lys is . Duxbury Press , Belmont CA . Laas, F.J., Forss, D.A. and Godfrey, M.E.R. 1985. Retention of brodifacoum in sheep t issues and excretion in faeces. New Zea land Journal of Agricultural Resea rch . 28:357-359. Lechev in, J .C. and Vigie, A. 1992. Which useful toxicological information can be drawn from studies on the hepatic fixation of anticoagulant rodenticides. Proc. Vert. Pest Conf. 15:204-207. Leek, J .B. and Park, B.K. 1981. A comparat ive study of the effects of Warfarin and brodifacoum on the relationship between vitamin K, metabol ism and clotting factor activity in Warfarin susceptible and Warfarin resistant rats. B iochemica l Pharmaco logy. 30:123-128. Link, K.P. 1945. The anticoagulant 3, 3-methyl bis-4-hydroxycoumarin. Fed . Proc. US:176-182. Linz, G.M., Davis, J.E., Engeman, R.M., Otis, D.L. and Avery, M.L. 1991. Estimating survival of bird ca rcasses in cattail marshes. Wildlife Soc iety Bulletin. 19:195-199. Lund M and Rasmussen A M . 1986. Secondary Poisoning Hazards to Stone Martens (Martes foina) fed bromadiolone poisoned mice. Nord. Vet-Med . 38:241-243. Lund, M. 1967. Warfarin resistance in Denmark. In: Report Intl. Conf. Rodents and Rodent ic ides, Paris. E P P O Publ . Ser. A 41:49-57. 127 Macfar lane, R.G. 1964. An enzyme cascade in the blood clotting mechan i sm and its function as a biochemical amplifier. Nature 202:498-499. Marzluff, J .M. and Ba lda, R.P. 1992. The P inyon Jay. Behavioral eco logy of a colonial cooperative Corvid. T& A D Poyser, London. McClymont, R.A., Fenton M. and Thompson, J.R. 1982. Identification of Cerv id t issues and hybridization by serum albumin. J . Wildl. Manage. 46(2):540-544. Mendenha l l V .M . and Pank, L.F. 1980. Secondary poisoning of owls by anticoagulant rodenticides. Wildl. Soc . Bull. 8(4):311-315. Merson, M.H., Byers, R.E., and Kaukeinen, D.E. 1984. Res idues of the rodenticide Brodi facoum in voles and raptors after orchard treatment. J . Wildl . Manage. 48(1):212-216. Miletitch, J .P. 1995. Prothrombin Time. L82-L84. In: Wi l l iams Hematology. 5th edition. E. Beutler, M.A. Lichtman, B.S. Co l l e r and T.J . Kipps, Editors. McGraw-Hi l l Inc. Health Profess ions Divisions. 1804 pp. Mineau, P. and Col l ins, B.T. 1988. Av ian mortality in agro-ecosystems 2. Methods of detect ion. B C P C Mono No.40. Environmental effects of Pest ic ides. 13-27. Mirkova, E. and Antov, G . 1983. Experimental evaluation of the risk of prenatal pathology under the effect of warfarin coumarin rodenticide. Hig Zdrav. 25(6):476-482. (Cited in I PCS 1995) Moore, N.W. (Editor) 1966. Pest ic ides in the environment and their effect on wildlife. J . of App l . Eco l . (Supplement). 34 pp. Moore, T.D., Spence , L.E. and Dugnolle, C.E. 1974. Identification of the dorsal guard hairs of some mammals of Wyoming. Hepworth, W.G . Editor. Wyoming G a m e and Fish Department. Moors, P.J. and Atk inson, I.A.E. 1984. Predation on seabirds by introduced animals, and factors affecting its severity. In: Status and Conservat ion of the World 's Seab i rds . J.P. Croxal l et al., Editors. I CBP Technica l Publ icat ion. 667-690. Mosterd, J . J . and Thi jssen, H.H.W. 1991. The long term effects of the rodenticide, brodifacoum, on blood coagulation and vitamin K metabol ism in rats. Br. J . Pharmaco l . 104:531-535. 128 Mount ME . 1988. Diagnosis and therapy of anticoagulant rodenticide intoxications. Veter inary Cl in ics of North Amer ica: Smal l Mamma l Pract ice. 18(1 ):115-131. Mount, M.E., Fe ldman, B.F., and Buffington, T. 1982. Vitamin K and its therapeutic importance. J A V M A . 180(11): 1354-1356. Murphy, M.J., Ray, A . C , and Bailey, E.M. 1989. A high performance liquid chromatography method for the detection of brodifacoum in serum. Vet. Hum. Toxico l . 31(3):228-230. Murphy, M.J., Ray, A . C , Woody, B. and Reagor, J . C 1985. Se rum brodi facoum concentrat ions and coagulopathic effects in anticoagulant poisoned dogs treated with vitamin K<. Toxicologist 5:88. Nagorsen, D.W. 1996. Opposums, Shrews and Moles of British Co lumbia . The mammals of British Co lumbia Vol.2. Royal British Co lumbia Museum Handbook. U B C Press Vancouver. 169 pp. Ne lson, L. Jr. and Clark, F.W. 1973. Correction for sprung traps in catch/effort calculat ions of trapping results. Journal of Mammalogy 54(1 ):295-298. Newby, S. and White, B.G. 1978. Brodifacoum: Adsorpt ion and desorption in soi ls measured under laboratory conditions. Has lemere, Surrey, Imperial Chemica l Industries Ltd., Plant Protection Division (Report No. T M J 1764B). (Cited in I PCS 1995). Newton, I., Davis, P.E., and Davis, J .E. 1982. Ravens and buzzards in relation to sheep farming and forestry in Wa les . J . App l . Eco l . 19:681-706. Newton, I., Wyll ie, I. and Freestone, P. 1990. Rodent ic ides in British Barn Owls. Environmental Pollution. 68:101-117. O'Reil ly, R.A. 1985. Anticoagulants, Antithrombotic, and Thrombolyt ic Drugs. C h . 58 1338-1359. In: Goodman and Gi lman's The Pharmacolog ica l Bas i s of Therapeut ics, 7 t h Edition. Editors: Goodman Gi lman, A., Goodman , L.S., Rai l , T.W., Murad, F. Macmil lan Publ ishing Company, New York. 1839 pp. Parmar, G., Bratt, H. Moore, R. and Batten, P.L. 1987. Ev idence for a common binding site in-vivo for the retention of anticoagulants in rat liver. Human Toxicology. 6(5):431-432. Poche , R.M. 1988. Rodent t issue residue and secondary hazard studies with bromadiolone. E P P O Bulletin. 18:323-330. 129 Radvany i A., Weaver , P., Massar i , C , Bird, D. and Broughton, E. 1988. Effects of ch lorophacinone on captive kestrels. Bull. Environ. Contam. Toxico l . 41:441-448. Rammel l C.G. , Hoogenboom J.J.L., Wil l iams J.M., and Bel l , J . 1984. Brodi facoum residues in target and non-target animals following rabbit poisoning trials. New Zea land Journal of Experimental Agriculture. 12:107-111. Rapaport, S.I. 1987. Introduction to hematology. 2nd edition. J B Lippincott Co . 403 pp. Record, C R . and Marsh, R.E. 1988. Rodentic ide residues in animal ca r casses and their re levance to secondary hazards. Proc. Vert. Pest Conf. 13:163-168. Redfern, R., Gi l l , J .E. and Hadler, M.R. 1976. Laboratory evaluation of W B A 8 1 1 9 as a rodenticide for use against Warfarin-resistant and non-resistant rats and mice. J . Hyg. 77:419-426. Resch , H., P ie tschmann, P., Krexner, E. and Wil lvonseder, R. 1991. Dec reased peripheral bone mineral content in patients under anticoagulant therapy with phenprocoumon. Eur. Heart J . 12:439-441. Rodway, M.S., Lemon, M.J.F. and Kaiser, G.W. 1990. British Co lumbia Seab i rd Co lony Inventory: Report #2- West Coas t Moresby Island. Canad ian Wildlife Serv ice Technica l Report Ser ies No. 65. Pacif ic and Yukon Reg ion. Rodway, M.S., Lemon, M.J.F. and Kaiser, G.W. 1988. British Co lumbia Seab i rd Co lony Inventory: Report #1- East Coas t Moresby Island. Canad ian Wildlife Serv ice Technica l Report Ser ies No. 50. Pacif ic and Yukon Reg ion. Ross , D.B., Roberts, N.L. and Cameron , D.M. 1978. The acute oral toxicity ( LD 5 0 ) of PP581 to the Mallard duck. Macclesf ie ld, Cheshire, Imperial Chemica l Industries Ltd, Central Toxicology Laboratory (Report No. ICI 122WL/78507). Ross , D.B., Roberts, N.L. and Cameron, D.M. 1977. The acute oral toxicity ( LD 5 0 ) of PP581 to the Ch icken. Imperial Chemica l Industries. Unpubl ished. Ross , D.B., Roberts, N.L. and Cameron, D.M. 1976. The acute oral toxicity ( LD 5 0 ) of PP581 to the Japanese Quai l . Imperial Chemica l Industries. Unpubl ished. S A S . 1995. J M P Statistical Package. S A S Institute Inc. Carey, North Caro l ina. 130 Savar ie , P.J., Hayes, D.J., McBr ide, R.T., and Roberts, J.D. 1979. Eff icacy and safety of d iphacinone as a predacide. In: Av ian and mammal ian Wildlife Toxicology. E.E. Kenaga , Editor. A m . Soc . Test. Mat. Spec . Tech . Publ . No. 693: 69-79. Scharde in , J.L. 1985. Anticoagulants. In: Chemica l ly induced birth defects. New York, Base l , Marcel Dekker Inc. pp. 89-100. (Cited in I PCS 1995). Searcey, M.T., Graves , C.B. and Olson, R.E. 1977. Isolation of a Warfarin binding protein from liver endoplasmic reticulum of Sprague-Dawley and Warfarin-resistant rats. Journal of Biological Chemistry. 252(18):6260-6267. Shirer, M. 1992. In poison's defence. Terra Nova 17:3. Smith, R.H., Cox, P.R. and Rampaud, M. 1990. Rodentic ide ecotoxicology: systems analys is and simulation. Proc. Vert. Pest Conf. 14:47-54. Stejskal, V., Kucerova, Z. and Zuska , J . 1994. Rodent bait containing brodi facoum as a developmental medium for a population of Ptinus tectus (Coleoptera, Ptinidae) yielding teratological individuals. Ochr. Rost l . 30(1 ):11-13. Stopforth, A. 1970. A study of coagulation mechan isms in domest ic ch ickens. J . Comp . Path. 80:525-533. Stutzenbaker, C D . , Brown, K. and Lobpries, D. 1984. Spec ia l Repor t : A n assessment of the accuracy of documenting waterfowl die-offs in a Texas coastal marsh. In: Workshop on Lead Po isoning in Wild Waterfowl, Sympos iums Proceedings. 88-95. Sul l ivan, T.P. 1990. Demographic responses of smal l mammal populations to a herbicide application in coastal coniferous forest: population density and resil iency. Can . J . Zoo l . 68:874-883. Suttie, J.W. 1980. The metabolic role of vitamin K. Federation Proceed ings. 39(10):2730-2735. Tahira, N., Dube, B. and Agrawal , G.P. 1977. Blood coagulat ion studies in s ome wild Indian birds: effect of different t issue thromboplastins. J . Comp . Path. 87:451-457. Taylor, D.P. 1984. The identification and detection of the rats in New Zea land and the eradication of ship rats on Tawhitinui Island. Unpubl ished dissertation for Dip loma in Parks and Recreat ion, Lincoln Col lege, Canterbury, New Zea land. 73 pp. (Cited in Eason and Spurr 1995). 131 Taylor, R.H. 1993. The Feasabi l i ty of Rat Eradication on. Langara Island, British Co lumbia , Canada . Unpubl ished report. 23 pp. Taylor, R.H. and Kaiser, G.W. 1993. Trip Report: Lucy and Cox Island at the North End of Haida Gwai i (Queen Charlotte lslands)-31 May to 6 June 1993. Unpubl ished report. 11 pp. Taylor, R.H. and Thomas, B.W. 1993. Rats eradicated from rugged breaksea island (170 ha), Fiordland, New Zea land. Biological Conservat ion 65:191-198. Taylor, R.H. and Thomas, B.W. 1989. Eradication of Norway Rats (Rattus Norvegicus) from Hawea Island, Fiordland, using Brodi facoum. New Zea land Journal of Ecology. 12:23-32. Thi jssen, H.H.W. and Baars, L.G.M. 1989. Microsomal Warfarin binding and vitamin K 2,3-epoxide reductase. B iochemical Pharmacology. 38(7):1115-1120. Thi jssen, H.H.W. and Baars, L.G.M. 1987. Hepatic uptake and storage of Warfar in. The relation with the target enzyme vitamin K 2,3-epoxide reductase. Journal of Pharmaco logy and Experimental Therapeut ics. 243(3): 1082-1088 Townsend M.G., Bunyan P.J., Odam, E.M., Stanley P.I., and Wardal l H.P. A s se s smen t of secondary poisoning hazard of Warfarin to least wease l s . J . Wildl. Manage. 48(2):628-632. Townsend, M.G., Fletcher, M.R., Odam E.M., and Stanley, P.I. 1981. A n assessment of the secondary poisoning hazard of Warfarin to tawny owls. J . Wildl . Manage. 45(1):242-247. V a n Haar lem, L.J.M., Knapen, M.H.J., Hamulyak, K. and Vermeer, C. 1988. Circulating osteocalc in during oral anticoagulant therapy. Thromb. Haemost . 60: 79-82. Walz , D.A., Kipfer, R.K. and Olson, R.E. 1975. Effects of vitamin K def ic iency Warfarin, and inhibitors of protein synthesis upon the p lasma levels of vitamin K-dependent clotting factors in the chick. J . Nutr. 105:972-981. Watson , J.W., Garrett, M.G. and Anthony, R.G. Foraging ecology of bald eag les in the Co lumbia River Estuary. J . Wildl. Manage. 55(3):492-499. 132 Worthing, C R . and Walker, S.B. 1987. The Pest ic ide Manua l . The British Crop Protection Counc i l , Thornton Heath, England. Lavenham Press . (Cited in Taylor 1993). 133 Appendix A. The Detection of Exposure to Brodifacoum in Japanese Quail through Plasma Brodifacoum Residue Analysis and Prothrombin Time Evaluation Introduction In 1995, the Canad ian Wildlife Serv ice attempted to eradicate introduced Norway rats from 3,300 hectare Langara Island at the north-western tip of British Co lumbia 's Queen Charlotte archipelago by use of the anticoagulant brodifacoum. The is land, however, is also home to breeding Bald Eag les, Common Ravens and other wildlife which prompted concerns of non-target spec ies poisoning. Bald Eag les , the top predator on Langara Island, were identified to be at risk from secondary poisoning from potentially scavenging/preying on rats and non-target spec ies exposed to brodifacoum. Bald Eag les were trapped and blood samp les drawn for prothrombin time (PT) determinations and p lasma brodifacoum residue analys is to determine if exposure occurred. The PT s were evaluated using the commercia l ly avai lable Coulter PT/Fibrinogen kit, a mammal ian (rabbit brain) thromboplast in. The validity of using a mammal ian source thromboplastin for avian PT determinations for brodifacoum exposed birds needs to be evaluated. One of the many tests to a s ses s the hemostatic system, the PT is used to indicate the relative risk of hemorrhaging of individuals suspected of anticoagulant exposure. The PT measures the extrinsic portion of the coagulat ion cascade (Brown 1988), and is an indicator of the vitamin K dependent clotting factor activity (Rapaport 1987). After whole blood is col lected and the ca lc ium is bound by sod ium 134 citrate to prevent coagulat ion, t issue thromboplastin is mixed with p lasma, and time to clotting is noted (Brown 1988). The PT is compared against a known normal and indicates the degree of clinical anticoagulation. A P T can be determined for avian spec ies using homologous or heterologous thromboplast ins. The PT which most accurately reflects the in-vivo system, however, is best estimated with a homologous thromboplastin. Didisheim et. a l . (1959) found that when homologous thromboplastin is used for PT determinations in avian p lasma, the prothrombin time is no longer than many mammal ian spec ies . However, the P T is prolonged when a mammal ian source thromboplastin is used for avian PT determinations (Kase 1978; Tahira et al. 1977; Didisheim et a l . 1959). The literature is unclear, but there is a suggest ion that factor VII is in low concentration or absent in avian blood (Walz et a l . 1975; Stopforth 1970), which may account for the increased PT if avian thromboplastin has no physiological need to bind with factor VII. Ka se (1978) found ev idence of factor VII in avian p lasma while Bellevi l le et al. (1982) demonstrated the mammal ian equivalent of factors V, VII, IX, and X in J apanese Quai l p lasma. The resulting prolonged prothrombin t imes, when using a mammal ian source thromboplastin in avian p lasma (and v ice versa), may reflect a difference between avian and mammal ian coagulat ion systems, or that they are less specif ic manifestations of the c lass specificity of these protein interactions (Kase et al. 1980). However, further studies by Bellevil le et a l . (1982) and Doerr and Hamilton (1981), have demonstrated a fundamental ly similar coagulat ion pathway as in the well studied human system, thus the dif ferences are 135 l ikely a result of the c lass specificity of the protein reactions. Therefore, the use of a mammal ian thromboplastin for evaluating avian PT may be adequate for compar ison purposes if the PT are reproducible between individuals. The objective of this experiment was to evaluate if brodifacoum residue could be detected in the p lasma of birds, and if the prothrombin test kit detect an increase in prothrombin time. The secondary objective was to a s se s s whether brodifacoum p lasma residue levels or prothrombin time change is a more sensit ive indicator of exposure to low levels of brodifacoum. 136 Materials and Methods 1.0 Plasma Brodifacoum Residues Eighteen male Japanese quail (Coturnixjaponica) were obtained from the U B C Quai l Genet i c Resource Centre. The birds were put on a 12L/12D light regime with water and food provided ad libitum. After a 6 day accl imatisation period the birds were randomly divided into six treatment groups of two with six as reserves. The experimental des ign was a 2 x 3 factorial, with two dose levels of brodifacoum and three time points for blood collection. The two dose levels were 0.7 mg/kg and 1.4 mg/kg (6% and 12% of the L D 5 0 respectively (Ross et al. 1976). The three time points for blood collection were 24 h, 5 d and 10 d post dos ing. After process ing the 24 hour groups, 2 of the 1.4 mg/kg group had died between 24-36 h post dosing, 1 by 3 d, and 1 by 4 d post dos ing. Necropsy results revealed mass ive local ised hemorrhaging on the top of the head and side of the neck. J apanese Quai l tend to jump in reaction to noise and possibly caus ing them to hit their heads on the ceil ing of the battery. Thus, the remaining 6 reserve quail were dosed at a lower level of 0.35 mg/kg and blood collected from 2 birds each at 24 h, 6 d and 10 d post dos ing. Trunk blood was collected into heparin coated test tubes and transferred into 1 ml cryovials for centrifuging. The p lasma was pipetted into 1 ml prelabel led cryovials and frozen at -20 C until shipment with the livers to the Department of Agriculture, State of Illinois, Veterinary Laboratory Serv ice for H P L C analys is (Murphy et al. 1989; Hunter 1983). The detection limit was 0.002 ppm. Al l birds were examined for s igns of internal bleeding and livers extracted and 137 frozen for residue analysis. The 0.7 and 0.35 mg/kg (instead of 1.4 mg/kg) dose groups data were treated as a 2 (dose levels) x 3 (time points) experiment for analys is. Ana lys is of the data was carried out using the JMP statistical package (SAS , 1995) with the following statistical model: Y i j k = u + D, + 7j + (DT)jj + E i j k where Y i j k = Res idue level, D, = the effect of the ith dose level, and TJ = effect of the jth time point of blood collection, (DT)jj = the two-way interaction between dose effect and time of blood collection, and E i j k = random error. The data was common log transformed in attempt to normalise the data and the analys is rerun using the above model . An A N O V A was used to analyse the 1.4, 0.7 and 0.35 mg/kg dose groups at 24 h post dos ing. 2.0 Prothrombin Time Validation Twenty seven, 4 month old male Japanese quail (Coturnix japonica) were obtained from the U B C Quai l Genet ic Resource Centre. The birds were put on a 12L/12D light regime with water and food provided ad libitum. After a 6 day accl imatisation period the birds were randomly divided into nine treatment groups of three. The experimental des ign was a 3 x 3 factorial, with three dose levels of brodi facoum and three time points post-exposure for prothrombin time evaluat ion. The three dose levels were 0.0 mg/kg (control), 0.7 mg/kg (a low and sub-lethal 138 dose), and 1.4 mg/kg (a high and potentially lethal dose as determined from a pilot study conducted earlier). The three time points for prothrombin time evaluation were 24 h, 72 h and 120 h post dosing (Stopforth, 1970). After process ing the 24h and 72 h treatment groups, it was obvious that the highest dose group were not showing signs of internal bleeding or morbidity after exposure. The 120h groups were therefore not processed, and the birds were given a 7 week rest period (to minimise any effects of prior exposure) and redosed at 13.5 mg/kg (the upper 9 5 % conf idence interval of the L D 5 0 (Ross et a l . 1976)). B lood was col lected from the jugular vein but trunk blood was col lected from those with which difficulties were encountered in bleeding. Blood col lected was immediately transferred into a test tube containing buffered sod ium citrate. Prothrombin t imes were measured within 6 h of collection using the Coulter P/T Fibr inogen kit (Lot # N1222295) for manual evaluation (tilt-tube method) (Brown, 1988). Al l birds were sacrif iced after blood collection and examined for s igns of internal bleeding. S ince blood col lected from decapitation showed a significantly shorter P T time than blood col lected via the jugular vein, the PT measurements were converted to prothrombin time ratio (PTR) to eliminate the bias due to collection method (Miletitch 1995). The P T R is the ratio of the dose group PT to the mean control PT (PTR= Samp le PT/ Mean Control PT). The data was square root (SQRT) transformed for analysis. Initial analys is showed that there was no difference in the P T R in the control groups of the first and 139 the added treatment, the data were therefore treated as a 4 (dose levels) x 2 (time points) experiment for analysis. Ana lys is of the data was carried out using the JMP statistical package ( S A S 1995) with the following statistical model: Y i j k = n + D| + 7j + (DT)ij + E i j k where Y i j k = (SQRT) P T R measured, D, = the effect of the ith dose level, and 7] = effect of the jth time point of blood collection, (DT)jj = the two-way interaction between dose effect and time of blood collection, and E i i k = random error. 140 Results 1.0 Plasma Brodifacoum Residues Log transformation had no effect on analysis. The data is presented using the arithmetic va lues. 1.1 Plasma Residue Levels There was no significant interaction between time and dose. 1.1.1 Time post Exposure There was no significant difference in brodifacoum p lasma residues at 24 h between the 1.4, 0.7, and 0.35 mg/kg dose groups. The p lasma residues decl ined significantly ( P O . 0 5 ) between d 1 (0.028 ± 0.005 ppm) and 5 (0.005 ± 0.005 ppm) d post dos ing, but not between d 5 and 10 (0.002 ± 0.005 ppm) (Figure A-1). 1.1.2 Effect of dose level There was a significant dose effect. The brodifacoum p lasma residue was significantly less (P<0.05) in the 0.35 mg/kg dose group (0.005 ± 0.004) than the 0.7 mg/kg group (0.018 ± 0.004) (Figure A-2). 1.2 Liver Residue Levels There was no significant interaction effect, or significant effect of t ime or dose on the level of residue between groups (Table A-1). 141 0.04 0.035 1 0.03 0.025 0.02 0.015 0.01 0.005 a -|— 24 132 Time Post Exposure (Hours) 240 Figure A -1 . Effect of time on brodifacoum p lasma residue concentration (ppm) after a single oral dose of brodifacoum at 0.35 and 0.7 mg/kg (means that do not share the same letter were significant at P<0.05) (n=4 at each time point). 142 g- 0.021 0.016 0.011 0.006 0.001 + 0.7 0.35 Dose Brodifacoum mg/kg Figure A-2. Effect of dose on brodifacoum residue concentration (ppm) over 10 d (* significant at P<0.05) (n=2 at each dose level). 143 Table A-1. Liver brodifacoum residue in Japanese Quai l after a single oral dose of brodifacoum (mean ±s.e., n=2 for each time/dose group). Days Post Dose (mg/kg) Dose 1.4 0.7 0.35 1 0.700 ±0.100 0.52 ± 0.110 0.487 ± 0.12 5 - 0.443 ±0.122 0.402 ± 0.003 10 - 0.354 ± 0.043 0.373 ± 0.037 1.3 Necropsy Results Quai l in the 1.4 mg/kg group showed hemorrhaging around the cranium and neck, and into the abdominal cavity. The 0.7 and 0.35 mg/kg groups showed no signs of hemorrhaging. 2.0 Prothrombin Time Validation There was no significant interaction between time post exposure and dose level. 2.1 Time post exposure One of the 4 birds dosed with 13.5 mg/kg brodifacoum showed s igns of morbidity by 48 h and the remaining were bled for PT evaluation before 72 h post exposure. These were included in the 72 h group for analysis. The P T R of birds bled at 24 h (1.11 ± 0.068 sec) was significantly (P<0.02) less than those bled at 48 or 72 h post exposure (1.37±0.062 sec) (Figure A-3). 2.2 Effect of dose level There was a significant (P<0.02) dose effect. The P T R for the control 144 1.45 1.4 g 1.35 03 CD 2 1.3 co _i c? 1.25 l— CL r 1.2 | 1-15 ro cl 1 - 1 1.05 1 24 48 Time Post Dose (hours) Figure A-3. Effect of time on the prothrombin time ratio (PTR) of J apanese Quai l after a single oral dose of brodifacoum (* significant at P<0.05) (n=9 at each time point). 145 (1.00+0.082 sec) was not significantly different from those of the 0.7 mg/kg group (1.17±0.088 sec). The P T R of the 13.5 mg/kg dose group (1.47+0.099 sec) was significantly longer than the control and 0.7 mg/kg groups but not the 1.4 mg/kg group (1.31 +0.099 sec). The 1.4 mg/kg and 0.7 mg/kg were not significantly different from each other (Figure A-4). 2.3 Necropsy Results The control, 0.7 mg/kg and the 1.4 mg/kg groups showed no s igns of hemorrhaging. The 13.5 mg/kg group showed s igns of hemorrhaging. One individual was found dead at 48 h post dosing with extensive hemorrhaging in the breast musc le as well as frank blood in the abdominal and thoracic cavity. Another individual which had not died at 48 h, showed signs of hemorrhaging on the left body wall of the abdomen. 146 1.8 1.7 co 1-6 1.5 1.4 o 1.3 o a: £ 1.2 co CO 1.1 1 1 0.9 n=5 ab n=5 be n=6 1 1.4 0.7 Dose Brodifacoum mg/kg n=7 13.5 Figure A-4. Effect of dose of the prothrombin time ratio (PTR) of J apanes quail after a single oral dose of brodifacoum (means that do not share the same letter were significantly different at P<0.05). 147 Discussion The results of the experiment indicate that the Coulter PT/Fibr inogen kit is able to detect an increase in the PT of avian p lasma after exposure to brodi facoum. However, the kit was unable to detect a low, sublethal exposure to brodi facoum. Brodi facoum residue was detected in the p lasma at all dose levels for up to 10 d. In the PT experiment both time and dose were significant, likely from the 1.4 and 13.5 mg/kg doses . The lack of significant interaction between time and dose may be a result of the low sample s ize in each group and/or variation between samples . The variability is likely due to the use of mammal ian thromboplastin for avian PT determinations and not the technique (Griminger 1986). Dorn and Mul ler (1965), using a rabbit brain thromboplastin, did not detect a significant increase in the PT s of ch icks until 6 d of feeding vitamin K deficient and 0.1% sul fonamide feed. Converse ly , the use of homologous thromboplastin detected a significant rise in the PT at 3 d. Thus, a mammal ian thromboplastin would not be as sensit ive as an avian thromboplastin for avian PT determinations. It would have taken a bigger dec rease in prothrombin levels before it can be detected by mammal ian thromboplast in. In other words, the PT/Fibrinogen kit may be useful to detect a high dose of brodifacoum but would not be sensit ive enough to detect a low dose exposure. The PT/Fibr inogen kit was unable to detect a low dose of brodifacoum that did not result in an increased PT or a high dose until the PT rose, 48-72 h after exposure. These results coincide with the accepted model that after an 148 anticoagulant binds in the liver and inhibits the recycling of vitamin K, enough time is required for the p lasma concentration of the vitamin K dependent clotting factors to decl ine to a point below which the PT increases (Rapaport 1987; Hoffman et a l . 1988). The increased P T R indicates a risk of bleeding but not that hemorrhaging will occur. The 1.4 mg/kg and 13.5 mg/kg dose groups had similar P T R s at 48/72 h although only the latter group showed any signs of bleeding. In the brodifacoum residue experiment, individuals in the 1.4 mg/kg group, lethally haemorrhaged apparently brought on by trauma. This indicates that trauma may be required to induce hemorrhaging in individuals that have suppressed prothrombin levels. The prothrombin levels may have been enough to stop minor bleeding such as from spontaneous haemorrhages, but not a "major" wound (Doerr and Hamilton 1981). The 13.5 mg/kg group prothrombin levels may have been suppressed to the point beyond which minor, spontaneous hemorrhaging could not be control led. The P T R of the 1.4 mg/kg group at 72 h suggests that these individuals were at risk of hemorrhaging although the conditions in the laboratory setting did not induce any detectable bleeding. Brodi facoum was rapidly removed from the blood stream between days 1 and 5 post dos ing fol lowed by a slower insignificant elimination to the end of the study at day 10. In mammal serum, brodifacoum follows an exponential decay, persisting for 2-3 w (Muphy et a l . 1985). Eason et al. (1996) reported that sub-lethal levels of brodi facoum (0.1 mg/kg) were detectable in p lasma of possums (Trichosurus 149 vulpecula) for 35 d after oral administration of a sub-lethal dose. The residues in quail liver indicate that brodifacoum is slowly el iminated. For a c losely related compound, f locoumafen, the elimination half-life for quail liver is >100 d (Huckle et al. 1989b). In mammals, brodifacoum is extremely slowly el iminated (Mosterd and Thi jssen 1991). Godfrey (1985) est imated the half-life of brodi facoum to be in excess of 150-200 d. Multiple long-term exposure to sub-lethal doses of brodifacoum have the potential to accumulate and suppress factor levels from a partially depressed state and may result in significant bleeding (Hoffman et al. 1988). A l though the test results are from a small sample s ize, they indicate that the use of a mammal ian derived thromboplastin used for the PT determination of bald eag les on Langara Island in 1995 was valid and would have detected a severe ly anticoagulated bird. However, a high PT would likely correspond to a positive detection of p lasma brodifacoum residue. When used in combinat ion, the p lasma brodifacoum residue and PT data can confirm exposure and assoc iated risk of hemorrhaging. However, p lasma brodifacoum residue analys is is a more sensit ive indicator of exposure in birds. 150 Appendix B. Brodifacoum Exposure in the Song Sparrow. Introduction Song sparrows were photographed at three rat ca rcasses placed to identify scavengers of rats (Chapter 4). They were apparently attracted to insects that began to consume the rat carcasses . In chapter 5, carrion insects were found to test positive for brodifacoum after consuming brodifacoum poisoned rats and pose a tertiary poisoning risk to non-target spec ies . The objective of this section was to determine if song sparrows were exposed to brodifacoum. Methods Song sparrows were col lected by shotgun after the intensive baiting period beginning mid August 1995. After selected morphological measurements were taken, the livers were removed and frozen for analys is. The livers were pooled across time or location for analysis by H P L C as descr ibed in Chapter 4. Results Brodi facoum residues were detected in sparrows col lected (Table A2.1). Se lec ted morphological measurements and collection locations can be found in table A2.2 . 151 Table B-1. Brodi facoum residues detected in Song Sparrows, Langara Island, 1995. T i ssue Brodi facoum Pool Sparrow Location Date Ana lysed (ppm) 1 1,4 North Eger ia Bay August 14 Liver ND 2 2 , 9 North Eger ia Bay August 16 Liver 0.643 3 6 , 7 South Eger ia Bay August 14/16 Liver ND 4 10, 11, 12 Lord Bight August 21 Liver 0.567 5 3, 5, 8, 13 Eger ia Bay/Lord August Body 0.058 Bight 14/16/21 Table B-2. Morphological measurements from Song Sparrows col lected on Langara Island, 1995. No Date Sex Wing Tail Hallux Bill Tarsus Tarsus Col lected Chord Length (cm) Depth Length Diameter (cm) (cm) (cm) (cm) (cm) 1 August 14 M 6.8 7.1 - 0.61 - -2 August 16 M 6.3 5.9 0.8 0.55 2.20 0.12 3 August 16 - - - - 0.59 - -4 August 14 M 6.6 7.6 0.78 0.60 2.34 0.18 5 August 14 - 6.2 7.0 - 0.70 - -6 August 14 F 7.0 6.9 0.86 0.65 2.72 0.20 7 August 16 M 6.4 5.8 - 0.62 2.48 0.17 8 August 16 - 6.3 7.2 0.81 0.54 2.36 0.16 9 August 16 M 6.9 7.6 0.86 0.72 2.31 0.18 10 August 21 M 7.2 5.7 0.82 0.62 2.64 0.16 11 August 21 M 7.0 6.5 0.80 0.70 2.50 0.16 12 August 21 M 6.6 5.6 0.79 0.61 2.44 0.16 13 August 21 - 7.0 6.0 0.74 - 2.47 0.16 152 Discussion Song sparrows were exposed to brodifacoum during the baiting campaign and were at risk of poisoning. It is unclear if the risk was from consuming carrion insects consuming brodifacoum poisoned rat ca rcasses or primarily from eating crumbs of bait that was found scattered around bait stations and along rat runs. In New Zea land, both granivorous and insectivorous birds were primarily po isoned after feeding on aerial distributed bait pellets intended for introduced rodents (Eason and Spurr 1995). Song sparrows are primarily insectivores, however, they are known to take seeds (Ehrlich et al. 1988). A s well, they may have preyed on snai ls or other invertebrates feeding on the bait in the stations. Therefore, song sparrows were at risk of primary, secondary and tertiary poisoning. Primary poisoning from the bait crumbs, secondary from the invertebrates feeding on the bait, and tertiary from carrion insects. The pooling of samples precludes quantifying the brodifacoum residue in individual ca rcasses as the brodifacoum may have originated from a single sparrow or all, likely diluting the concentration. However, it can be conc luded that sparrows were at risk of brodifacoum poisoning and the primary, secondary and tertiary poisoning risk should be further investigated in future eradication operat ions. 153 Appendix C. Potential Sub-Lethal and Long Term Effects of Brodifacoum Exposure There is no data from the literature on the long term effects of brodifacoum or the other anticoagulants on target or non-target spec ies. This appendix identifies possib le sub-lethal effects and potential long term effects of brodifacoum exposure to non-target spec ies . The anticoagulant rodenticides are known to suppress clotting factor levels, affect bone mass, reproductively toxic, and are teratogenic with possible deve lopment effects. The suppress ion of the clotting factor levels was the mechan i sm focused on in this thesis, and the mode of action is d i scussed in Chapter 2. However, there are long term implications of sub-lethal or repeated exposure to brodifacoum in that an increasingly smal ler dose would potentiate the anticoagulant effect because of the already suppressed level of clotting factors. A normal PT may be obtained from human p lasma with between 30-100% of normal vitamin K dependent clotting factor concentrat ions (Hoffman et al. 1988). Suppress ion of the clotting factor concentration below the threshold would result in an increased PT and possibly uncontrol lable hemorrhaging. A significant decrease in the levels of the vitamin K dependent clotting factors lasted 43 days after ingestion of brodifacoum (Hoffman et a l . 1988). Bone mass may be compromised by sub-lethal or long term exposure to oral ant icoagulants. Bone contains the vitamin K dependent protein osteocalc in which is part of the bone matrix, and levels in the blood may be used as an indicator of active bone deposit ion. The oral anticoagulants (phenprocoumon and acenocoumaro l ) reduce the activity of osteocalc in, which is activated by a vitamin K dependent reaction 154 (Van Haar lem et a l . 1988). Patients on long term oral anticoagulant therapy showed significantly lower bone mass than controls (Fiore et a l . 1990; Res ch et a l . 1991). Furthermore, it was found that a poor vitamin K status was assoc iated with a high urinary ca lc ium loss (Knapen et al. 1993). Therefore, young individuals with growing bones and female birds and mammals with high calc ium demands in the breeding season are at greatest risk of sub-lethal effects of brodifacoum exposure. The coumarin anticoagulants have been shown to be embryotoxic and teratogenic to rats. In rats, warfarin induced increased rates of embryolethality, hemorrhage and gross structural malformations (internal hydrocephalus and anomal ies of skeletal ossification) (Mirkova and Antov 1983). Similarly, indiscriminate use of brodifacoum in fields was related to increased inc idences of abortions and hemorrhage in sheep and goats in Egypt (Feinsod et a l . 1986). In contrast, other laboratory studies have not demonstrated reproductive toxicity or teratogenicity of brodi facoum to rats or rabbits at various dose levels (Hodge et a l . 1980a, 1980b, 1980c). Ant icoagulants have developmental impacts on the human fetus. In humans, there are two types of anticoagulant induced defects, depending on the time of administration of the anticoagulant: fetal warfarin syndrome and fetal wastage (Anonymous 1976). The most consistent feature of fetal warfarin syndrome is nasal hypoplas ia leading to respiratory difficulty. Fetal wastage results a lso in central nervous system anomal ies. Other common features includes bone abnormalit ies of the axial and appendicular skeleton ( IPCS 1995), opthalmological malformations 155 leading to bl indness, developmental delay, low birth weight, premature birth, mental retardation, and ear anomal ies (Schardein 1985). A l though the above summary indicates that brodifacoum may have potential long term effects from sub-lethal exposure to non-target spec ies, more work in these areas are required to confirm potential impacts. However, this summary identifies potential areas of investigation that may be conducted both in laboratory and field exper iments. Clearly, the mortality endpoint as a result of the exposure to brodifacoum, the main focus of this thesis, is not the only non-target spec ies impact that could be measured during future island restoration projects. 156 Appendix D. Environmental Aspects of Brodifacoum - Transport, Distribution and Transformation. Air, Water and Soil Air Anticoagulant rodenticides have low volatility and increased levels in the air are unlikely ( IPCS 1995). Water Brodi facoum is slightly soluble in water (<10 mg/l at 20 °C, pH 7). The vapor pressure is <0.13 m P a at 25°C ( IPCS 1995). It is a weak acid and does not readily form water-soluble salts (Worthing and Walker 1987). However, it is known to be toxic to fish. The brodifacoum 96-h L C 5 0 to rainbow trout is 0.051 mg/l (Hill et al. 1976). This is equivalent to 1 bait block (20 g bait with brodifacoum at 0.005%) per 20 I of water that could be cons idered a toxic hazard to f ish. Soil The adsorption and desorption of 1 4 C-brod i facoum in laboratory condit ions has been investigated. Brodi facoum binds very strongly to soil particles and equil ibria is establ ished fairly rapidly with larger watensoi l ratios despite very low brodifacoum water solubility (Newby and White 1978). Binding increases with greater organic matter in the soil (ICI 1984). Once bound to the soil particles, 1 4 C-Brod i f a coum is effectively immobile and desorption is very s low (Jackson and Hall 1992; Newby and White 1978). Less than 2% of brodifacoum, added to 157 soil (0.6 and 6.0 kg/ha) with a pH from 4.3-7.1, organic matter from 6.8-72.1% and clay content from 5-19%, leached more than 2 cm (ICI 1984). This suggests that mobility of brodifacoum would be restricted to erosion processes, traveling with soil particles. Once bound to soil particles, s low microbial degradation reduces brodifacoum to C 0 2 and water (Taylor 1993; Shirer 1992). The half life of brodifacoum in soil ranges between 12-25 weeks, depending on soil condit ions (ICI 1984). Hall and Priestley (1992) monitored the metabol ism of 1 4 C -Brodi facoum in soil under aerobic conditions for 52 weeks. A mean of 35.8% of radioactivity recovered was 1 4 C 0 2 . Rad io labe led brodifacoum was the major rad io labe led component in the soil extracts. A half-life of brodifacoum in soil was calculated to be 157 days. No information on the abiotic degradation of brodifacoum is avai lable. However, abiotic degradation of related second-generat ion anticoagulants have been descr ibed. Bromadio lone degrades rapidly on exposure to sunlight, with a half-life of 2.1 h ( IPCS 1995). The photolytic half life of d i fenacoum at pH 5, 7, and 9 over 24 hours was calculated to be 3.3, 8.1 and 7.3 h (Hall et a l . 1992). The hydrolytic half-life of brodifacoum was found to be in excess of 30 days, but no precise estimation was made (Jackson et al. 1991). These data indicate that brodifacoum is a lipophilic compound relatively resistant to breakdown in soil. Any brodifacoum that is not consumed by target or non-target organisms and that is left in the environment will be present and 158 potentially avai lable for transport into the ecosystem long after the bait has been removed from the stations. Biological Retention of Brodifacoum Biologically, brodifacoum is a lipophilic anticoagulant rodenticide relatively resistant to metabol ism. Most (75%) of a 25 mg/kg dose of brodifacoum given to rats was retained principally in the liver, pancreas and the sal ivary glands, at ten days post dosing when the study ended (Godfrey 1985). The biological half-life was found to be 150-200 days (Godfrey 1985). Parmar et al. (1987) est imated the liver brodifacoum half-life to be 130 days in rats. Brodi facoum was detected in sheep liver 128 days post dosing at 0.2 and 2 mg/kg (Laas et a l . 1985). Similarly, sheep consuming a single sub-lethal dose of 2 mg brodifacoum/kg showed 2 mg/kg in the liver 4 months after dosing (Rammel l et al. 1984). These data indicate that sub-lethally exposed mammals and likely birds, pose a long term poisoning risk to spec ies preying on them. Long term retention of brodifacoum may also lead to accumulat ion of brodifacoum initiating the anticoagulant effect leading to hemorrhaging and death. 159 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0087971/manifest

Comment

Related Items