Home Page | Interdisciplinary Centre for Conservation Science
Quantifying the interactions between humans and endemic pitvipers (Bothrops caribbaeus) in Saint LuciaPhotograph: K. BreachKatherine BreachSupervisors: E.J. Milner- Gulland and Matthew MortonA thesis submitted in partial fulfilment of the requirements for the degree of Master of Science and the Diploma of Imperial College LondonContentsList of acronyms……………………………………………………………………………………………………………….…4Abstract……………………………………………………………………………………………………………………………...5Acknowledgements……………………………………………………………………………………………………….….61. Introduction…………………………………………………………………………………………………….……….……71.1 Human-snake conflict and the conservation of Bothrops caribbaeus……………………………71.2 Problems to be addressed…………………………………………………………………………………….……..71.3 Aims and objectives……………………………………………………………………………………………………..82. Background…………………………………………………………………………………………………………..………92.1 The genus Bothrops (Serpentes, Viperidae)………………………………………………………….……..92.2 Human-wildlife conflict……………………………………………………………………………………………….122.3 Snakebite- a global health problem…………………………………………………………………………….142.4 Related research…………………………………………………………………………………………………………152.4.1 Community surveys of snakebite……………………………………………………………………….…….152.4.2 Studies of human-snake interactions………………………………………………………………….…..162.4.3 The use of Geographic Information Systems (GIS) to create risk maps………………..….173. Methods………………………………………………………………………………………………………………….…193.1 Methodology………………………………………………………………………………………………………….….193.1.1 General methodological approach………………………………………………………………………...193.1.2 Methodological links to related research……………………………………………………………..203.1.3 Logistical limitations…………………………………………………………………………………………….203.2 Application of methodology…………………………………………………………………………………….213.2.1 General community survey…………………………………………………………………………………..213.2.1.1 Selection of community survey locations……………………………………………………………213.2.1.2 Application of the community survey………………………………………………………………..213.2.2 Wildlife professional…………………………………………………………………………………………233.2.3 Bite victims……………………………………………………………………………………………………....243.2.4 Hospital records………………………………………………………………………………………………..253.2.5 Mapping sightings and bite locations………………………………………………………………..253.2.6 Historical data search and mapping of historical range…………………………………….263.3 Statistical analysis………………………………………………………………………………………………..274. Results…………………………………………………………………………………………………………………...284.1 General sample characteristics…………………………………………………………………………….284.2 Outcomes of snakebite…………………………………………………………………………………………314.3 Identifying correlates of snakebite risk………………………………………………………………..334.4 Treatment preferences………………………………………………………………………………………..384.4.1 Medical versus non-medical treatment…………………………………………………………….384.4.2 First aid preferences………………………………………………………………………………………….404.5 Attitudes and behaviour towards B. caribbaeus…………………………………………………..434.5.1 Local attitudes to B. caribbaeus………………………………………………………………………..434.5.2 Local attitudes towards hypothetical conservation protection………………………….454.5.3 Abilities in identifying B. caribbaeus…………………………………………………………………464.5.4 Respondent behaviour upon encountering B. caribbaeus………………………………..474.6 Examining geographic variation in human-B. caribbaeus conflict………………………..494.7 Investigating historical evidence of past species range……………………………………….575. Discussion…………………………………………………………………………………………………………….595.1 Decreasing snakebite risk……………………………………………………………………………………595.2 Addressing treatment preferences, attitudes and behaviours…………………………….605.3 Distribution maps……………………………………………………………………………………………….615.4 Other potential conflict mitigation methods……………………………………………………….615.4.1 Translocation……………………………………………………………………………………………………615.4.2 Humane snake traps……………………………………………………………………………………….625.4.3 Barrier methods……………………………………………………………………………………………..635.4.4 Deterring snakes from community areas……………………………………………………….635.4.5 Respecting B. caribbaeus in its natural habitat………………………………………………635.5 Limitations………………………………………………………………………………………………………..645.5.1 General community survey……………………………………………………………………………645.5.2 Bite victims……………………………………………………………………………………………………645.5.3 Hospital records…………………………………………………………………………………………….655.5.4 Distribution maps…………………………………………………………………………………………655.6 Further research……………………………………………………………………………………………..655.7 Conclusions…………………………………………………………………………………………………….66References…………………………………………………………………………………………………………..67Appendix……………………………………………………………………………………………………………..71Appendix 1: General community survey questionnaire………………………………………..71Appendix2: Hospital records form………………………………………………………………………..81List of acronymsHWC Human- wildlife conflictWWF World Wide Fund for NatureGIS Geographic Information SystemsNDVI Normalised Difference Vegetation IndexANBG Australian National Botanic GardenPIT Passive Integrated TransponderESRI Environmental Systems Research InstituteDWCT Durrell Wildlife Conservation TrustSLFD Saint Lucia Forestry DepartmentSDT Short Distance TranslocationLDT Long Distance TranslocationUSGS United States Geological Survey AbstractHuman-snake conflict on the West Indian island of Saint Lucia is a problem for both human safety and for the conservation of Bothrops caribbaeus- the endemic pitviper involved. Snakebite by B. caribbaeus can be a traumatising but rarely deadly event for the bite victim. Despite the rarity of snake bite incidents, B. caribbaeus is a much feared and despised animal, and is often killed on sight. The problem of human-snake conflict in St. Lucia has so far never been addressed, and previous to this study no work has been carried out to identify any potential conservation needs of B. caribbaeus. The aim of this investigation is to contribute to the understanding of the conservation needs of B. caribbaeus through an analysis of human-snake conflict. Correlates of snake bite risk were investigated in order to identify risk factors that need to be considered in order to help prevent snake bite. The outcomes of snakebite in terms of the injuries sustained were examined in order to assess the impact that snakebite has on human health. Local treatment preferences were investigated in order to identify potential need for awareness raising on treatment and first aid matters. In order to quantify local feelings towards B. caribbaeus, attitudes and behaviour towards the snake were analysed. The geographic variation in human- B. caribbaeus conflict was examined, and key conflict zones identified. Finally, past species range was investigated in order to examine any changes in distribution and identify possible conservation need.Methods included snowball sampling of bite victims, the search of hospital records, community based surveys, interviews with wildlife professionals and GIS mapping of conflict hotspot areas, inferred current B. caribbaeus distribution and historical distribution.It was found that snake bite is a fairly rare event, and although snakebite related mortality does occur, the majority of bite victims make a full recovery. Being unaware of B. caribbaeus (i.e. not seeing it) proved to be the most significant factor in explaining differences between people who were bitten and people who merely encountered the snake and were not bitten. The majority of people would and do visit the hospital in the event of B. caribbaeus bite. However, a large proportion of people also rely on herbal remedies and traditional healers, and many people carry out damaging first aid techniques such as tying tourniquets. Attitudinal surveys revealed that B. caribbaeus is very unpopular with St. Lucian people, and around half of encounters result in human-induced snake mortality. Mapping of historical range revealed that there has been a decline in species range, which highlights the need for conservation action. However, before physical conservation strategies can be implemented, it is essential that conservation education and awareness raising is implemented in order to change current attitudes and gain the support of local people. AcknowledgementsI would like to thank Matthew Morton and the St. Lucia Forestry Department for all the help and support provided for this project, without which this research would not have been possible. Thank you to Ellie, Erica, Jenni and Haydn for being such excellent company in St. Lucia. I am particularly grateful to my field assistant Curtis for all his help in the field, as I could not have managed to collect data without him. I would also like to thank Stephen Lesmond, Tim Jn. Baptiste and Morisson for all their help whilst in St. Lucia. I would like to say a particularly big thank you to Robert Devaux, for being a great friend and mentor, and for allowing me to undertake the historical research. Without him this part of the project would not have taken place. I would also like to thank Pam and Betty for such enjoyable weekends and lovely lunches! Thank you to Dr. Didier, Dr. King and Dr. Rambally for their input into this project, and for allowing the search of the hospital records. A big thank you to Janice Gaspard who put in so much time and effort searching these records for me. I would also like to thank Dr. David Warrell and Dr. Jenny Daltry for their advice and guidance, and for allowing me to see my first Fer de Lance! I would also like to thank Aiden Keene and Marcus Rowcliffe for help and guidance on the statistics. I would particularly like to say a big thank you to my mum, my dad, my brother and the rest of my family for supporting me all the way through this project, and to James Lambert for being so supportive and understanding. I would like to thank Markus Ultch for letting me use his photograph. Finally, I would like to thank my supervisor E.J. Milner- Gulland for being so incredibly supportive and understanding. This dissertation would not have been completed without such intensive help!1. Introduction1.1 Human snake conflict and the conservation of Bothrops caribbaeusHuman-snake conflict on the West Indian island of Saint Lucia is a problem for both human safety and for the conservation of Bothrops caribbaeus- the endemic pitviper involved. B. caribbaeus is reported to be responsible for around 12 people being bitten each year, which can be a traumatising but rarely deadly event for the bite victim. Despite the rarity of snake bite incidents, B. caribbaeus is a much feared and despised animal, and is often killed on sight. The species is currently not protected, and its conservation status is unknown.Conflict between humans and B. caribbaeus in St. Lucia has been a long-standing problem, so much so that in 1870 the government introduced a bounty on snake heads, paying local people for every B. caribbaeus head brought to a police station. It is also documented that the Indian Mongoose (Hepestes javanicus) was brought to St. Lucia solely for the purpose of controlling snakes, unlike on other Caribbean islands (R. Devaux, in preparation). The problem of human-snake conflict in St. Lucia has so far never been addressed, and previous to this study no work has been carried out to identify any potential conservation needs of B. caribbaeus. 1.2 Problems to be addressedA variety of problems must be addressed in order to gain a better understanding of human-snake conflict and identify ways to reduce it. The factors affecting the risk of snakebite in St. Lucia are unknown. It is important that these risk factors are understood, as prevention strategies must be implemented to reduce bite risk and potentially aid in decreasing conflict. Although this is one component in conflict reduction, another key component is to understand local attitudes towards the species, and how local people react to encounters. To reduce conflict, negative attitudes must be reversed, and understanding the feelings and behaviours of local people will be necessary before this can happen. It is important to identify areas of the island where conflict is most frequent, as mitigation management strategies will need to focus on these locations. Many people are unaware of the distribution of B. caribbaeus across the island (the current range has not been formally identified), and this knowledge is important for conservation of the species as well as for identifying areas where precautions against snake bite are needed. The conservation need of B. caribbaeus is also unknown, and it is important that a greater understanding of this is reached so as to aid in conservation planning and prioritisation.Finally, from a human health perspective it is important to address current problems and issues surrounding treatment and first aid preferences associated with bite incidents, as it is thought that a proportion of bite victims do not go to hospital, and may carry out damaging first aid techniques. Understanding current practices will aid in identifying the need for education regarding how best to act if bitten by B. caribbaeus.1.3 Aims and ObjectivesThe aim of this investigation is to contribute to the understanding of the conservation needs of B. caribbaeus through an analysis of human-snake conflict. This will be achieved by addressing the following objectives:Investigate the outcomes of snakebite in terms of the physical injuries sustainedIdentify correlates of snakebite riskInvestigate local treatment preferencesInvestigate local attitudes towards B. caribbaeus and human behaviour upon encountering a snakeExamine geographic variation in human- B. caribbaeus conflict through mapping the locations of snake bites and sightings, and identifying key conflict zonesInvestigate historical evidence of past species range through mapping of locations of where B. caribbaeus had previously been foundSuggest methods to reduce human-snake conflict and address conservation needs2. Background2.1 The genus Bothrops (Serpentes, Viperidae)The genus Bothrops is made up of approximately 37 species of pitvipers (Campbell and Lamar, 2004), and are commonly referred to as “Fer de Lance” or “Lance head” vipers. Members of the genus Bothrops can be found in Central and South America, and on the Caribbean islands of Saint Lucia (B. caribbaeus), Martinique (B.lanceolatus) and Trinidad (B.atrox).This genus includes species responsible for the majority of snakebite incidents in the New World (Sasa, Wasko and Lamar, in press). There is morphological diversity and great variation in natural history between species in this genus. For example, juveniles of the widely distribute B. asper feed mainly on ectotherms (notably frogs and lizards). When this species reaches maturity, there is a dietary shift towards rodents and other small mammals, birds and even large anurans (Sasa, Wasko and Lamar, in press). In comparison to the generalist diet of B. asper, the adult diet of B. alcatraz is made up almost entirely of ectotherms, particularly centipedes and lizards (Marques, Martins and Sazima, 2002), while 85% of the diet of B. insularis is made up of migrant passerine birds (Martins et al, 2001). These more specialist diets would seem to be environmental adaptations as a result of the absence of small mammals on the islands that they inhabit (Marques, Martins and Sazima, 2002). The great diversity between species owes to the fact that Bothrops was the first group of pitvipers to reach the continent of South America, and have therefore had a significant amount of time for adaptive radiation. Bothrops caribbaeus is endemic to the island of Saint Lucia, and is known locally (in the Creole speaking community) as “Sépan” (figure 1). There is popular local belief that B. caribbaeus was introduced to St. Lucia by Europeans as a cruel method of control against enslaved people, or that it had been brought to the island as a weapon of war by the Carib Indians against the resident Arawaks. However, B. caribbaeus has inhabited the island long before man, and DNA sequencing by Wüster et al (2002) has shown that B. caribbaeus of St. Lucia, and the endemic B. lanceolatus of Martinique have not recently diverged from Latin American Bothrops species, but split from the Latin American Bothrops asper-atrox complex around 4.2- 8.9 million years ago, dating back to the late Miocene or earliest Pliocene. B. caribbaeus and B.lanceolatus therefore represent relatively ancient, independent evolutionary lineages. Morphological data has led to the hypothesis of a “stepping-stone” colonisation, with St. Lucia being colonised first by dispersal from the mainland of Latin America, followed by a further dispersal event to Martinique from St. Lucia (Wüster et al, 2002). These two Lesser Antillean species form a monophyletic group, which forms a sister clade with the Bothrops asper-atrox complex (Wüster et al, 2002).Figure 1: Bothrops caribbaeus (photograph by K. Breach)There is relatively little work published on B. caribbaeus in the scientific literature, especially regarding the species ecology. Sasa, Wasko and Lamar (in press) commented that the knowledge gap regarding the natural history of many Bothrops species may be explained by the general decrease in herpetological natural history studies, coupled with the fact that species of Bothrops are cryptic, are often found in unfavourable habitats for humans, are usually nocturnal and have a reputation as being dangerous.B. caribbaeus is semi-arboreal, and Lazell (1964) reported collecting a specimen from a height of 6 meters. However, this species is most often found on the ground, and Lazell (1964) reported that the majority of the specimens he collected where found amongst piles of rocks of coconut husks. The average size is documented to be around 90-120cm (3- 4 feet), although there has been anecdotal reports of the species reaching 8 feet or larger. Figure 2 demonstrates the potential size that this species can grow to. After conducting a study on B. caribbaeus, Lazell (1964) reported that the body of one individual (which had been beaten to death) measured “two meters without the head”. Schwartz and Henderson (1991) documented that examination of stomach contents (of an unspecified number of specimens) found only the remains of introduced Mus and Rattus species. However, there has been anecdotal evidence of B. caribbaeus preying not only on rodents but on other small mammals as well as birds. Long (1974) reported that even invertebrates such as slugs have been found in the stomachs of B. caribbaeus. The species has been described as an avid predator of introduced rats, and Lazell (1964) details that rats caused virtually no damage to Cocoa crops in areas where B. caribbaeus is abundant, compared to up to 70% rodent-induced crop damage in areas where B. caribbaeus is not found. However, this again seems to be anecdotal evidence, and in the present day the wide-spread use of rodenticide and other chemicals will probably alter this finding. There was a sighting (witnessed by several people) of a male and female B.caribbaeus mating in March 2009, and although anecdotal, many locals say that the height of the breeding season for this species is around May. B. caribbaeus is viviparous, and Lazell (1964) documented the removal of 26 unborn young from a gravid female. Schwartz and Henderson (1991) recorded the removal of 37 developing embryos from another female. B. caribbaeus is widely thought to be nocturnal and therefore particularly active at night. When sighted during the day, it is said that B.caribbaeus is usually resting on the ground in a coiled position, or occasionally found sunning itself. Figure 2: A large adult B. caribbaeus that has just been killed (Photograph by Markus Ultch)2.2 Human- wildlife conflictHuman-wildlife conflict (HWC) refers to confrontation between people and wild animals, and the resulting negative outcomes and impacts on humans and their resources, and to wild animals and their habitats. It is brought on when the requirements of people and wildlife overlap, creating costs to both parties (DiStefano, 2005). HWC can affect human health, safety and welfare, the sustainability and security of livelihoods, and can even affect governments, industries and national economies (WWF report, 2008). Wildlife can damage and destroy people’s property, crops and livestock, can spread disease, and even kill and injure people. Often the people most affected by HWC are among the most impoverished in the world (WWF report, 2008). HWC also has major environmental impacts, and human-induced injury and mortality caused by conflict is a serious threat to the survival of many species worldwide. Not only is the species directly involved in the conflict affected (as animals are often killed and injured by people in retaliation), but the outcomes of conflict may affect whole ecosystems through trophic cascades and even habitat destruction (Woodroffe, Thirgood and Rabinowitz, 2005).Snakes, especially venomous ones, are maligned throughout much of the world. As found in one study conducted in Australia by Whitaker and Shine (2000), many people living in the presence of snakes have the attitude that “the only good snake is a dead snake”. The ability of some snake species to seriously injure and kill people has earned them a fearful reputation. Figure 3 and 4 show examples of the injuries sustained as a result of B. caribbaeus bite. From an anthropocentric view, human injury and loss of life is probably the most negative and extreme form of HWC. From an individual viewpoint, snakebite can be debilitating and even deadly. Disability caused by snakebite can lead to loss of livelihood, and other negative social impacts associated with venomous snakes include loss of livestock and pets, restriction of travel and fear. At a larger scale, snakebite can have an economical impact, for example when considering the effects of disability of victims from an economically productive age group and the high cost of antivenom. According to Kasturiratne et al (2008), when analysing snakebite incidences globally, it was found that the majority of people bitten by snakes fell in this economically productive age bracket. Theakston and Warrell (2000) reported that up to 70% of hospital beds in north-eastern Nigeria can be taken up by snakebite victims at certain times of the year. Lalloo et al (1995) reported that certain regions of Papua New Guinea have among the highest rates of snakebite in the world, and Cheng and Winkel (2001) found that a single vile of polyvalent antivenom in Papua New Guinea costs the equivalent of half the annual gross domestic product per capita. Treating snake envenomations in such situations can have huge financial implications. From findings such as these it can be seen that there is potential for the effects of snakebite to have high economic impacts. Figure 3: This man needed his forearm Figure 4: This man suffered permanent amputated after a B. caribbaeus bite scarring and nerve damage (Photograph by K. Breach) (Photograph by K. Breach) On the other hand HWC can have greatly negative impacts on snakes and their environments. Many snakes, including unoffending individuals are deliberately and indiscriminately killed by humans in retaliation or because of fear for their own lives, or for the lives of their family or animals. Figure 2 and figure 5 demonstrate this problem. Even harmless, non-venomous species may be persecuted, probably due to a lack of knowledge on species identification, as well as the fact that snakes in general are often disliked. The number of snakes killed by humans greatly exceeds the number of humans killed by snakes (Whitaker and Shine; 2000). Figure 5: B. caribbaeus killed as a result of human-snake conflict (photograph by K. Breach)The threats posed by venomous snakes may be very real, but it has been documented that sometimes people can perceive these threats to be greater than they are in reality (e.g. Whitaker and Shine; 2000), and therefore kill snakes according to the perceived rather than the realistic threat. For example, with regards to the venomous Eastern Brownsnakes (Pseudonaja textilis) and the venomous Common Blacksnakes (Pseudechis porphyriacus) in Australia, it was found that humans where 20 times more likely to approach a snake, and 100 times more likely to attack a snake, compared to the likelihood of a snake approaching or attacking a person (Whitaker and Shine; 2000). In this study it was found that around a third of snakes seen by people where killed. Defensive displays by the two species where often mistaken for attacks, and people overestimated aggressive responses, as non-aggressive responses from the majority of encounters resulted in the snake being undetected by the passing human (Whitaker and Shine; 2000). Snakes play an important role within an ecosystem and are among the top predators in their habitat, influencing the populations of their prey species. Significant reduction or removal of snakes can severely upset the finely tuned balance of the food chain. Snakes often predate on (and therefore help control) rodents, many of which are introduced pests that spread disease, damage crops and other resources, and have detrimental impacts on native flora and fauna. For example, one single rattlesnake can eat between 20-30 rodents per year. Removal of these top predators can therefore lead to an increase in rodent pests, as well as having other affects on the ecosystem. 2.3 Snakebite- a global health problemEnvenoming by B. lanceolatus and B. caribbaeus seems to produce a unique syndrome that is different from that produced by other Bothrops species. While envenoming by Latin American species in the B.asper-atrox complex are characterised by a defibrination and bleeding syndrome, envenomings by B. lanceolatus and B.caribbaeus produce a unique systemic thrombotic syndrome (Wüster et al, 2002). A World Health Organisation funded project by Kasturiratne et al (2008) found that across the world, snakebite is responsible for a minimum of 421,000 envenomations and 20,000 deaths each year. These numbers could be as high as 1,841,000 envenomations, and 94,000 deaths annually. Kasturiratne et al (2008) calculated that between 1.2 million- 5.5 million snakebites (including dry bites and bites from non-venomous species) may occur globally per year, based on the evidence that only around one in four snakebites results in envenomation. The majority of these cases are concentrated in South and Southeast Asia, sub-Saharan Africa and Central and South America. The country with the highest estimated number of snake envenomations and deaths is India, with nearly 11,000 deaths annually. The Caribbean (the area within which my study site lies) was estimated to have between 1,098-8,039 envenomings per year, and between 107-1,161 deaths per year, as a result of snakebite from several species. Although the estimations may be crude, and to the authors own admittance “far from perfect”, Kasturiratne et al’s (2008) study is the most scientifically robust and replicable of the studies into global snakebite statistics to date. Despite the extrapolation and estimation, it is evident that snake envenomation is truly a worldwide health problem.2.4 Related research2.4.1 Community surveys of snakebiteMost investigations and reports of snakebite incidence are based on hospital-based data. However, this information is unlikely to truly reflect actual snake bite incidence, particularly in developing countries where many people can not easily access health facilities, or prefer to use traditional medicine. Community based surveys are therefore beneficial as they yield more reliable data that more truly reflects snakebite incidence in an area.A number of studies have used community-based surveys as a method of investigating snakebite incidence (e.g. Hati et al, 1992; Watt et al, 1987). Sharma et al (2004) conducted a community based study in south-eastern Nepal, aiming to evaluate the impacts of snake bite and determine the risk factors linked to fatal outcomes. It was estimated that 10,000 people represented by approximately 1,700 households needed to be surveyed in order to obtain an adequate sample of snakebite victims, given that the expected annual incidence of snakebite was 300/ 100,000. A total of 1,817 households were sampled, and 143 bite cases were recorded, with 75 of these cases showing evidence that envenomation occurred. Of the 143 reported cases, 20 cases resulted in mortality. Based on the household surveys, the annual rates for total snake bite incidence, envenomation incidence, and bite related mortality incidence were calculated. Risk factors associated with bite related mortality were also assessed. To identify significant explanatory variables associated with bite related mortality, comparisons were made between victims with fatal and non-fatal outcomes. The most significant risk factors included being bitten inside the house, while sleeping, and/ or being bitten between 12am and 6am, visiting a traditional healer before seeking medical help, and having a lack of transport to medical facilities or delay in obtaining transportation. The study by Sharma et al (2004) presents some valuable information, and the large sample size (1,817 households) means that the results can be generalised to the area with more confidence compared to if hospital records or a small sample size had been used. The snake bite incidence rate observed was very high (1,162/100,000) compared with many other regions of the world. The probability of finding an adequate number of bite victims through random sampling is fairly likely when snake bite incidence is very high. However, in regions of the world where snakebite incidence is fairly low, the likelihood of finding snakebite victims through a random sampling method is low, and the chances of collecting enough data on bite victims to make meaningful comparisons (for example when comparing those with fatal verses non-fatal outcomes) would be difficult. Therefore, for studies that are carrying out an investigation that is dependent on information from bite victims, and are operating in areas of lower snakebite incidence, a non-probabilistic sampling method may be more suitable. There are various non-probability sampling methods that may be employed to detect hard to reach or hidden subpopulations, for example the use of respondent-driven sampling (e.g. Salganik and Heckathorn, 2004) or the use of snowball sampling. When the target group are not necessarily part of an interlinked network (i.e. they are not part of the same social group), snowball sampling (where existing respondents recruit future respondents from among their acquaintances), is more appropriate than respondent-driven sampling. Many studies have used this method to sample hard to reach target groups, for example Avico et al (1998) used a snowball sampling method in order to locate cocaine users, McNamara (1994) used the technique to investigate male prostitution, and Pollack and Schlitz (1988) used snowball sampling to identify AIDS sufferers. The main limitations with this method is that the sample is not representative of the wider population, and that the sampling frame cannot be identified (unlike when using respondent-driven sampling).2.4.2 Studies of human-snake interactionsIn order to mitigate human- snake conflict, it is important to understand how people react to the snakes that they encounter. This way the level of conflict and the effects that it has on snakes can be better understood.While investigating sources of mortality for venomous Eastern Brownsnakes (Pseudonaja textilis) and Common Blacksnakes (Pseudechis porphyriacus) in Murrumbidgee Irrigation Area (MIA) in Australia, Whitaker and Shine (2000) conducted a survey to ascertain people’s attitudes and behaviour towards snakes in the area. This survey was conducted by distributing 1,000 questionnaires to both rural and urban residents around MIA. Questionnaires where distributed to school children to take home, as well as being left in public buildings. The questionnaire sought information on snake encounters within the 12 months prior to the survey, and was designed to gather information about when, where, how and why people responded to encountering a venomous snake. A total of 138 questionnaires where returned. Results revealed that people responded in the same way to both venomous species, despite one species being far deadlier than the other (Whitaker and Shine, 2000). Around 50% of encountered snakes were approached by the observer, and about a third of encountered snakes were killed. Women who encountered a snake were found to be more likely to kill the snake then men. The questionnaires found that many people attack snakes because they perceive the snake to be aggressive, and therefore fear them, and are worried about the potential danger posed to themselves, their children or their pets. However, further study of both species responses to encounters suggested that the majority of people largely overestimate the levels of aggression shown by this snake in reality. 2.4.3 The use of Geographic Information Systems (GIS) to create risk mapsThere are numerous studies that adopt a GIS approach to mapping tropical disease risk, for example the global mapping of malaria risk (Hay and Snow, 2006) or the prediction of schistosomiasis infection risk in Africa (Brooker et al, 2002). However, there has been very little previous work that employs GIS mapping as a method of determining areas of snake bite or snake encounter risk.Molesworth et al (2003) used ArcView 3.1 GIS to map epidemiological and environmental data in order to identify potentially high snakebite risk areas in Northern Ghana and Nigeria, so as to aid in health care planning. The study focused on mapping snake bite incidence caused by the West African Carpet Viper (Echis ocellatus), a species that is responsible for the majority of snakebite related mortality in the study area (Warrell and Arnett, 1976). A risk map to reflect the probability of high potential snakebite incidence was compiled using a) snakebite incidence rates associated with different health facilities and their 15km catchment areas, b) Average NDVI ( normalised difference vegetation index) and c) Savannah regions under 100m elevation, where envenomation due to Echis ocellatus is thought to be restricted to. Molesworth et al’s (2003) study seems to be one of (if not the) first pieces of published research to use GIS to map snake bite risk, albeit at the level of the health facilities and not at an individual level. The use of environmental data such as NDVI to investigate whether these variables have any affect on snakebite incidence is also very useful in understanding factors that influence snakebite, as well as providing some insight into the species ecology. However, this study has many limitations. By the authors own admission, the risk map can only indicate the probability of high snakebite incidence at the level of the health facility, and not individual snakebite risk (Molesworth et al, 2003) as the data collected was at too coarse a scale. However, there are complications with this in itself. The methods used for estimating incidence rates may introduce many errors, for example using the arbitrary 15km catchment area around the health facility as a representation of the area where the bites had occurred. Also, using only data obtained from health facilities does not truly reflect snakebite incidence, as facilities are unevenly distributed and many people in developing countries will preferentially use herbal remedies and visit a traditional healer. This could be rectified by carrying out investigations into incidence rate at an individual or community level across the study area, therefore data would not represent only those communities with health facilities. Molesworth et al (2003) highlight the fact that 6 of the health facilities shared overlapping catchment areas, which would make the assumption of independence invalid (Molesworth et al, 2003), and add further error to calculations of snakebite incidence for each facility. The above limitations associated with using the annual number of snakebite cases for each health facility as a representation of local snakebite incidence rates therefore seems deeply flawed. The epidemiological data were just one of three sets of data used to compile the risk map to reflect the probability of high potential snakebite incidence. Average NDVI was also used, as it was significant in explaining patterns of high snakebite incidence rates. However, if there are errors in the calculated incidence rates, then perhaps this result is also misleading. Further study will be needed to investigate this. Finally, the third key component was the map data on Savannah regions were snake bite was thought to be restricted to. However, while evidence from 30 years ago seems to suggest that snakes were abundant in this area and bites were frequent (Warrell and Arnett, 1976), there is no suggestion that snakes are found exclusively in savannah regions. More robust evidence will be needed before assuming the species range and using this as a component of risk mapping. However, Molesworth et al (2003) identify the need for further research, and caution against the use of the risk map as a definitive tool in health care decision making. Douglass (2008) used ArcGIS Version 9 to map the observed ranges of individual Eastern Brown snakes (Pseudonaja textilis) in the 90ha Australian National Botanic Gardens (ANBG). The Eastern Brown snake is responsible for the majority of Australian snake-bite related deaths. Mapping was carried out in order to identify snake “hotspots” where snakes are most likely to be encountered, as these areas may pose a risk to people. Snakes found in ANBG were safely captured, implanted with passive integrated transponders (PITs) and marked with non-toxic paint in order to permanently identify each individual. Locations of sightings, as well as information on capture and release locations were used to map individual snake activity ranges using minimum convex polygons in ArcGIS. Identified “hotspot” areas were used to improve snake management activities in ANGB, in order to avoid human-snake interaction and conflict. Douglass (2008) suggests that understanding and mapping snake activity ranges is crucial for conservation.In this example, GIS mapping of individual home ranges, and the identification of potential conflict hotspots proved feasible and successful in terms of conserving the species and protecting human health within the botanical garden. However, these methods may not be suitable for addressing human- snake conflict in other cases, especially where there is a lack of capacity to carry out such methods. Also, over a much larger area, with no physical boundaries and many snakes (for example, across regions of St. Lucia), this method would be very difficult to implement, and investigating individual home ranges when there may be hundreds of individuals would be almost impossible. 3. Methods3.1 MethodologyThis study uses a combination of different methods in order to investigate human-snake conflict in St. Lucia3.1.1 General methodological approach Figure 6 outlines the general methodological approach used in this study, showing which data sources were used to answer the different research questions.Identify bite risk, local attitudes and behaviours, and treatment preferencesComparison of two separate bite samplesBite victimsWildlife professionalsHospital recordsHistorical publicationsCommunity survey = Data source = Research question= used to map conflict areas/ current range= used to identify bite victims through snowball sampling= used to map historicalrange = used to investigate biterisk variables, attitudes, behaviours and treatment preferences= used to examine differences between two samples of bite victimsMap potential conflict areas and inferred current rangeMap historical rangeFigure 6: The general methodological approach used to investigate various research questions 3.1.2 Methodological links to related researchThis study uses a combination of different methods in order to investigate human-snake conflict in St. Lucia. In order to gain a better understanding of the impacts and outcomes of snake bite on humans, and determine the risk factors linked to it, snake bite incidence was researched at the community level. The importance of a community based investigation for researching snake bite and risk factors was highlighted in the study by Sharma et al (2004) who’s findings gave a more true reflection of snakebite incidence than if hospital records alone had been analysed, as a proportion of bite victims do not attend hospital. However, as snakebite events are uncommon in St. Lucia, a different sampling method was used to find this rare target group. Snowball sampling was used in order to sample bite victims at the community level. The usefulness of this method for locating a hard to reach target group has been demonstrated in many studies for example that carried out by Pollack and Schlitz (1988) for sampling Aids sufferers. In my study, hospital records were also consulted in order to compare the two bite victim samples.In order to understand conflict and the potential impacts on the “problem” animal species involved, it is important to understand local attitudes to the species, and how people react when the animal is encountered. This was carried out by Whitaker and Shine (2001) through the distribution of questionnaire surveys regarding snakes. In my study, local attitudes and responses were investigated as part of a general community based survey. Risk mapping is important for identifying potential conflict areas where risk mitigation strategies may need to be implemented. Both Douglass (2008) and Molesworth et al (2003) used GIS to map high snake conflict risk areas. Here, two very different scales were used; Douglass (2008) mapped risk at a very fine scale, based on individual snakes, while Molesworth et al (2003) used a very coarse scale, based at the level of the health facilities. The method of GIS mapping used in this investigation is at a more intermediate scale, based on sightings and bite incident locations across the island. GIS mapping was also used to map historical range, so that comparisons could be made with current distribution to identify any potential change in range size.3.1.3 Logistical limitationsThe main logistical limitations of this investigation involved time limitations, for example a limited number of hours could be spent at each settlement whilst conducting the general community survey (affecting sample size), and this time would have been during working hours only. Also a limited amount of time could be spent searching hospital and historical records. Another logistical limitation concerned the hospital records, as these were poorly organised which impacted the ease at which these could be searched. Other limitations included being unable to involve cases of snakebite mortality (due to limited data being available from the community level study), and being unable to investigate people who had encountered B. caribbaeus unawares, resulting in no conflict. 3.2 Application of methodology3.2.1 General community survey3.2.1.1 Selection of community survey locationsA probability sampling approach to community based surveying was taken, based on the methods of M. Morton (unpublished study). Twenty communities were stratified by Forestry Administration Range (Saint Lucia Forestry Department (SLFD)), then sampled by human population density and randomly sampled within that (figure 7). This was carried out as follows:Strata were allocated in ArcView GIS 3.2 (ESRI), using the five Forestry Administration Ranges; Northern Range, Dennery Range, Quilesse Range, Soufriere Range and Millet Range and census data per “settlement” (area including ≥ 1 communities) from the 2002 national census (St. Lucia Statistics Department). Settlement areas were classified into four quantile classes of human population density (people per hectare; 2002 census) as follows:1- 102104- 341342- 1,0551,056- 20,357This produced 20 strata. Using the ArcView Animal Movement Extension, one point was randomly allocated within each stratum. The random points within each area were then visually inspected against a 1:25,000 Ordinance Survey Map plus a St. Lucia Statistics Department map of houses. The points were then moved to the nearest mapped settlement within that area and stratum. If there was no mapped settlement within a “settlement” area that a random point fell in, a second random point was allocated within that stratum.3.2.1.2 Application of the community surveyPrior to the application of the main community survey, a one day pilot study was undertaken involving twenty local people around the plantation areas of St. Joseph and Errard Estates, Dennery. The pilot study was undertaken to test the ease of delivery and ease of understanding of the prepared questionnaire, so that adaptive changes could be made to it if necessary. This exercise was also useful in training both myself and my field assistant for the full scale community based survey. It was found from this that the printed questionnaire did not need to be changed, however, the verbal wording of the questions when asked to the respondent had to be modified in order to make it clear that responses should be in regards to only the last B. caribbaeus sighting made.Figure 7: Map to show the 20 communities sampled using the five forestry ranges and four quantile classes of human population density (Reproduced with kind permission by Matthew Morton)For the main community survey, each selected location was surveyed by both myself and a specifically trained fieldworker from the St. Lucia Forestry Department, who was able to speak the local Creole language.Seven to twelve people within each of the 20 communities were sampled opportunistically, and were interviewed using a standardised questionnaire (Appendix 1). The respondent’s age and gender were recorded, but no personally identifying details were collected. Core questions involved details about the respondents most recent B. caribbaeus sighting (if sighted at all), such as where the incidence had occurred, the time of day, the season, the activity the respondent was undertaking at the time, the density of any vegetation that they were in, what protection against snakebite they were wearing (if any) etc. Respondents were asked whether or not they had killed this particular snake and whether or not they had ever been bitten by B. caribbaeus in their lifetime. There were also a series of attitudinal questions relating to B. caribbaeus and questions regarding behavioural responses to a hypothetical snake encounter or snake bite. Finally, each respondent was asked if they knew any bite victims, and the names and locations of any such contacts where recorded. It should be noted that the information for a maximum of one B. caribbaeus sighting was recorded per respondent.The majority of questions were closed questions involving checklists (e.g. for the time of day, the habitat, the level of protective clothing, what the respondent would do if they encountered a snake etc). This structure was used as it produces information that is easy to quantify and analyse. However, there were also a couple of open ended questions that followed closed questions (for example, the reasons behind people’s attitudes). These questions allowed for elaboration and produced a greater depth of qualitative information. An attitude scale was used to investigate respondent’s feelings towards B. caribbaeus. The scale used included six options; strongly dislike, dislike, neutral (neither like nor dislike), like, strongly like and don’t know. The location of any B.caribbaeus sighting reported was identified and recorded using gridded Ordinance Survey map sheets of St. Lucia, at a scale of 1:25,000. Each map sheet represented a 20km2 grid (5km x 4km), with each kilometre square referencing the British West Indies Grid (Transverse Mercator, Clarke 1880). Each km2 on the gridded map sheet was numbered, and the numbered square that the point location fell into was recorded.3.2.2 Wildlife professionalsThe majority of local St. Lucians do not often spend time in the rainforested interior of the island, therefore most B. caribbaeus sightings from the general community survey will be biased towards areas surrounding communities and areas of agriculture. For the purpose of developing understanding of B. caribbaeus distribution, a number of “wildlife professionals” were non-randomly sampled from each of the five Forestry Administration Ranges (SLFD) in order to acquire further information on current range, especially in under represented areas such as rainforest. These people were considered to be reliable professionals who had had much contact with B. caribbaeus in the wild (especially in rainforest habitats under represented by sightings from the general community survey). This sample of people was recommended for interview by either Durrell Wildlife Conservation Trust (DWCT) or SLFD, and mainly involved experienced SLFD range workers.Each respondent was encouraged to give information for each B. caribbaeus sighting that they could recall, as the information obtained was to be used purely for understanding current B. caribbaeus distribution, and for constructing an ArcView GIS map to reflect these potential conflict areas. The location of each sighting was identified and recorded using the same method as for the general community survey.Finally, each wildlife professional was asked if they knew any B. caribbaeus bite victims, the names and locations of these people where recorded. 3.2.3 Bite victimsThe incidence of snakebite in St. Lucia is fairly low compared to many other regions of the world, and therefore bite victims are likely to be a hard to reach target group. A snowball sampling method was used in order to reach this rare subpopulation. All wildlife professionals and respondents from the general community survey were asked whether they knew anybody who had been bitten by B. caribbaeus. Any information regarding a bite victim was noted, particularly the name and where this person could be located. Bite victims found using this method where also asked if they could identify any other bite victims. Therefore, names of potential bite victim respondents were identified through wildlife professionals, respondents from the general community survey, and through other bite victim respondents.Bite victims seemed to be clustered into certain areas, therefore names of potential bite victim respondents were organised into lists according to location, and these locations were visited in turn. In order to interview as many identified bite victims as possible, most locations were visited on several occasions, and at different times of the day. Local residents at each visited location were often consulted to aid in finding the individuals identified by the snowball sample, especially as many of the names collected through sampling often referred to the same person, as having several nick-names seems to be common tradition.When each potential bite victim respondent was located, verbal consent to be interviewed was obtained, as well as confirmation that this person had been bitten by B. caribbaeus. Each consenting bite victim was interviewed using a standardised questionnaire. The respondent’s full name, age and gender were recorded. Core questions involved details about the bite incidence, such as where the incidence had occurred, the time of day, the season, the activity the respondent was undertaking at the time, the density of any vegetation that they were in, what protection against snakebite they were wearing (if any) etc. Questions on variables associated with the bite incident (such as those listed above) were the same questions as those used in the general community survey questionnaire to address variables associated with sighting incidents that did not result in a bite. Other questions in the standardised questionnaire for bite victims involved questions relating to what treatment was sought after being bitten, the injuries they sustained, and the time it took for them to recover. Finally each respondent was asked if they knew any other bite victims, and the names and locations of these people were recorded. The location of each bite incident was identified and recorded using the same gridded map sheets in the same way as was used in the community survey and with wildlife professionals. Only information relating to one bite was recorded per person- in the event that a person had been bitten more than once, only details on the most recent bite was used, unless the information could not be clearly recalled, in which case details on the previous bite were noted. 3.2.4 Hospital recordsIn order to compare bite victim respondents with a separate sample of St. Lucian bite victims, hospital records (Accident and Emergency admission logbooks, and the Morbidity Database) from Victoria Hospital, Castries, St. Lucia where systematically searched by a professional Health Information Management Technologist, and data from all discovered snake bite patients was recorded on a standardised form (see Appendix 2). The information documented included the patient’s age, gender, date of bite, date of admission, geographical location of incidence, bodily location of bite, outcome, duration of hospital treatment etc. A total of 35 hours was spent searching the records, which equated to 20,000 individual records being searched. The search was carried out as systematically as possible, starting from the most recent admissions (June 2009) and working backwards.Although there are several hospitals on the island, Victoria Hospital located in the capital city is the main hospital in St. Lucia, and all snakebite cases are transferred here from other hospitals. Due to confidentiality reasons and strict Ministry of Health (St. Lucia Government) policy, only staff currently working with the hospital records could have access to these, and no data released had any personally identifying information attached to it. 3.2.5 Mapping of sightings and bite locationsThe km2 grid cells from the Ordinance Survey map sheets that were identified as encompassing the location of B. caribbaeus bites and sightings (identified from the snowball sample of bite victims, the general community survey and from interviews with wildlife professionals) were mapped using ArcView GIS 3.2 (ESRI: Environmental Systems Research Institute). Individual maps were created to show B. caribbaeus sightings from the community survey, sightings from wildlife professionals and the bite locations from sampled bite victims. A final map reflecting the locations identified from all three samples was constructed to show the probable distribution of B. caribbaeus. From this estimated distribution map, potential conflict areas can be inferred, and “hotspot” areas of potential high conflict can be identified from any clustering of bite incidence locations or clustering of sighting locations. Each location is represented as a 1km x 1km square on the map, as identified by using the gridded map sheets.3.2.6 Historical data search and mapping of historical rangeIn order to delimit the historical range of B. caribbaeus, the following historical documents and records, located at either the St. Lucia National Archives, the St. Lucia Research Centre, or the Roman Catholic Chancery Archives were searched systematically for references to B.caribbaeus (apart from the Police Daily Occurrence Registers which (due to time constraints) were randomly sampled between the years of 1900-1944):St. Lucia Gazette (years 1891-1964)Annual Report on the St. Lucia Blue Books (years 1845-1964)The Voice of St. Lucia (years 1885- 1952)Snake head Purchase Registers (years 1927- 1940)St. Lucia Civil Status Registers (years 1752-1850)Police Daily Occurrence Registers (randomly sampled between 1900-1944)Many of these large record books were published annually, and some publications (e.g. The Voice of St. Lucia national newspaper) were published several times per week. These records and documents had been meticulously searched and bookmarked by local historian and naturalist Robert J. Devaux (OBE) prior to the start of this study, and therefore the references to B. caribbaeus related pages and articles in the above publications were more easily found. With the assistance of Mr. Devaux, each reference to this snake species was looked up in the relevant publication, and any significant details were noted. References within the publications were searched in chronological order: starting with the oldest and working forward. In particular, information was sought regarding the dates and locations of reported snake bites, snake sightings, and of snake heads collected for the bounty offered at the time.Historical data were only used for mapping historical range if sufficient details on the location were given and if information was deemed as reliable (one-off, purely anecdotal sightings were dismissed - evidence of B. caribbaeus at a location was required either by documentation of B. caribbaeus heads being collected at the location, a bite occurring, evidence of this species being killed in the area or multiple reports of sightings in an area that were later investigated and confirmed by the government or the press). This verification was required to make sure information used was as reliable as possible, and not a one-off sighting that may have been a misidentified Boa constrictor (Boa constrictor orophias).In order to create a map of historical range from the information collected, each location was identified and recorded using the gridded Ordinance Survey map sheets of St. Lucia. ArcView GIS 3.2 (ESRI) was used to map these locations to create a representation of the estimated B.caribbaeus historical range. Each location is represented as a 1km x 1km square on the map, as identified by using the gridded map sheets. The locations were mapped at this scale in order to remain consistent and comparable with the maps created from recent sightings and bite locations. The map is colour-coded to reflect the decade the location refers to. The most recent locations are from the 1940s, as there was no mention of B. caribbaeus in the literature after this.3.3 Statistical analysisAs the data were all categorical and non-parametric, Chi squared tests and Fishers Exact tests were used to analyse the data, and p<0.05 was selected as the level of statistical significance throughout. All graphical representations of data were created using Microsoft Office Excel 2007, and the analysis were carried out in R version 2.9.1.4. Results4.1 General sample characteristicsFor the general community survey, 169 respondents were interviewed. Most respondents were males (56%) and the most commonly represented age group were those between 18-39 years old (Table 1). From the sample, 67 respondents (40%) inhabited settlements that fell within the estimated current range of B. caribbaeus, and 102 respondents (60%) inhabited settlements that did not. A total of 47 (28%) of respondents had seen a live B. caribbaeus in the wild (i.e. as apposed to in a zoo or one that had been brought to a location for entertainment). One respondent (0.59%) had been bitten by B.caribbaeus. MaleFemaleTotal≥60 years15 (9%)8 (5%)23 (14%)40-59 years28 (17%)22 (13%)50 (30%)18-39 years43 (25%)43 (25%)86 (51%)≤ 17 years8 (5%)2 (1%)10 (6%)Total94 (56%)75 (44%)169 (100%)Table 1: Number of males and females from each age group interviewed in the community survey (percentages rounded to the nearest integer)From the sample of bite victims, a total of 47 people were interviewed. Most bite victims were male (83%) (Table 2), and most respondents were of working age (18-59 years) (78%). The majority of bite victims sampled (51%) were bitten between 2000 and 2009 (Table 3). Eleven bite victims (23%) had been bitten by B. caribbaeus on more than one occasion (although only one bite per person was included in the study). MaleFemaleTotal≥60 years4 (9%)0 (0%)4 (9%)40-59 years15 (32%)3 (6%)18 (38%)18-39 years16 (34%) 3 (6%)19 (40%)≤ 17 years2 (4%)2 (4%)4 (9%)Unknown2 (4%)0 (0%)2 (4%)Total39 (83%)8 (17%)47 (100%)Table 2: Number of male and female bite victims from each age group (percentages rounded to the nearest integer)YearNumber of bite victims2000-200924 (51%)1990-19995 (11%)1980-19895 (11%)1970-19797 (15%)1960-19693 (6%)1940-19492 (4%)Unknown1 (2%)Table 3: The proportions of bite victims bitten in each decadeA total of 22 snake bite patients (24 bite incidences, as two patients had been bitten twice) were identified by the systematic search of 20,000 hospital records. All patients were male, and most were of working age when they were bitten (87%) (table 4).MaleFemaleTotal≥60 years2 (8%)0 (0%)2 (8%)40-59 years14 (58%)0 (0%)14 (58%)18-39 years7 (29%)0 (0%) 7 (29%)≤ 17 years1 (4%)0 (0%)1 (4%)Total24 (100%)0 (0%)24 (100%)Table 4: Number of males and females from each age group identified by hospital records (percentages rounded to the nearest integer)Comparisons were made between bite victims identified from the snowball sample and snake-bite patients identified from the hospital records, in order to examine any differences. The majority of bite victims (66% of sampled bite victims and 87% from hospital records) in both samples were males of working age (between 18-59 years). No female patients were identified from hospital records, although they made up 17% of bite victims from the snowball sample. Table 5 shows a comparison between the outcomes of bite victims from the two samples.Snowball sampleHospital recordsFull recovery72%58%Permanent scarring/ nerve damage15%12%Amputation of bitten limb4%0%Currently still recovering9%0%Death0%4%Outcome not known0%25%Table 5: Outcomes of snakebite for bite victims from the snowball sample compared to those identified by hospital records.It can be seen that the majority of bite victims made a full recovery, (only 1 patient (4%) identified from the hospital records died) however these figures can only be used for comparison and only limited inferences can be made, because in 25% of cases the outcomes of snake bite were not recorded in hospital records. Any potential biases that there may be in the snowball sample of bite victims cannot be identified through comparison with the hospital records of bite patients, as a proportion of people bitten by snakes do not go to hospital. Table 6 shows the bodily locations of snakebite for sampled bite victims and those identified in hospital records.Snowball sampleHospital recordsLower leg42%50%Foot31%25%Hand11%21%Lower arm7%0%Upper leg7%0%Buttocks2%0%Data deficient4%4%Sample size4724Table 6: Bodily locations of snakebite for bite victim respondents and hospital patients from medical records.Both sets of data show that the most common place to be bitten is on the lower leg, followed by the foot, and then the hand. A Fisher’s Exact test revealed no significant difference between the bodily locations of snakebite between the two samples (p=0.51).Although the bite victims occupation was not recorded in the snowball sample, it is interesting to learn from the hospital records that the most common occupation of snakebite patients was that of a farmer (46% of patients) followed by labourer (25% of patients).Table 7 presents the times taken for those bite victim respondents who sought medical treatment to reach Victoria Hospital (the only hospital that treats snakebite). These data were unfortunately not available from hospital records, and percentages are calculated from the number of bite victims in the snowball sample who visited hospital, not of the total number of bite victims interviewed. Thirty four (72%) of bite victim respondents went to hospital at some point.Time taken to reach Victoria HospitalPercentage of respondents who visited hospital< 1 hour22%1<2 hours31%2<3 hours28%3<4 hours6%5<6 hours3%The following day6%3 days later3%Table 7: The time taken by bite victims that sought medical treatment to get to Victoria Hospital.It can be seen that the majority of bite victims reached hospital within 3 hours of being bitten (81%). The main reasons for any delay included visiting a traditional healer first, visiting a health centre or another hospital that can not treat snakebite first, attempting to treat themselves first using herbal remedies, believing that the injury was not serious, or not even realising that it was a snake that had bitten them. 4.2 Outcomes of snakebiteOf the 47 bite victims, 72% made a full recovery after being bitten. At the time of surveying (June-July 2009), 9% of bite victims were still recovering from their recent bite. A total of 15% of respondents had been left with permanent scarring and/ or nerve damage, and 4% of respondents had had their bitten limbs amputated (this involved two respondents who each had an arm amputated). The decade when bite victims where bitten was a significant predictor of whether they had made a full recovery (p=0.04; Figure 8; table 8). Bite victims who were recorded as still recovering were not included in this analysis. Surviving bite victims who did not make a full recovery where bitten between 1980 and 2009, with the 1990s having the highest percentage of surviving bite victims who did not make a full recovery (80% of respondents that were bitten in that decade). The likely reason for these results is that in recent decades, a larger proportion of bite victims are visiting hospital, and the permanent injuries recorded here (amputated limbs, permanent scarring and nerve damage) are caused mainly by respondents with severe envenoming that had to undergo life-saving surgical interventions . Those in earlier decades who had such acute poisoning would have possibly not survived without such medical intervention. As this study did not include cases of mortality, differences between those who survived and those who did not could not be assessed.Figure 8: The physical outcomes of snakebite per decade for the 47 bite victims sampledExplanatory VariableSignificanceStatistical test usedAge when bittenNot significant(p=0.41)Fisher’s Exact testDecade bittenSignificant(p=0.04)Fisher’s Exact testWhether or not went to hospitalSignificant(p=0.04)Fisher’s Exact testThe time taken to reach hospitalNot significant(p=0.11)Fisher’s Exact testSize of snakeSignificant(p=0.03)Fisher’s Exact testTable 8: Explanatory variables tested for significance in explaining the differences between those bite victims who fully recovered, and those who did not.Whether or not the bite victim went to hospital was statistically significantly different between those who made a full recovery and those who survived but did not make a full recovery (p=0.04). All sampled bite victims that did not make a full recovery did visit the hospital. However, as discussed above, the reason for this is that loss of limb and permanent scarring/ nerve damage was caused by potentially life-saving surgery in hospital. Without this, the respondent would have quite probably not survived. Bite victims who did not visit hospital will have obtained no such injuries, and may have possible died therefore would not be included in this study.The size of the snake was a significant factor affecting whether people made a full recovery and those who survived but did not make a full recovery (p=0.03; Figure 9). Bite victims bitten by smaller snakes all made a full recovery, unlike those bitten by larger snakes.Figure 9: the physical outcomes of bite victims bitten by differently sized snakes4.3 Identifying correlates of snake-bite risk In order to identify variables that may explain the differences between encountering a snake and being bitten and encountering a snake and not being bitten, data from bite victims (47 respondents) was compared with data from those from the community survey who had encountered B. caribbaeus in the wild and had not been bitten (also 47 respondents). There were 14 explanatory variables that were tested for statistical significance; these are listed in table 9.Explanatory variableSignificanceStatistical test usedWhether or not the respondent saw the snake before any potential bite incidentSignificant (X2=81.72, d.f.=1, p<0.01)Chi Squared testThe density of ground vegetation where the respondent was walkingSignificant(X2= 23.05, d.f. = 2, p<0.01)Chi Squared testWhether or not the respondent tried to kill the snake before any potential bite incidentSignificant(X2=11.52, d.f.=1, p<0.01)Chi Squared testWhether or not the respondent trod on the snakeSignificant(X2=20.12, d.f.=1, p<0.01)Chi Squared testThe time of year (season)Non-significant(X2=0.37, d.f.=1, p=0.54)Chi Squared testThe time of dayNon-significant(X2=4.09, d.f.=2, p=0.13)Chi Squared testWhether or not the respondent was walking on a pathNon-significant(X2=0.16, d.f.=1, p=0.69)Chi Squared testWhether or not the encounter took place on or near a road.Non-significant(X2=2.88, d.f.=1, p=0.09)Chi Squared testWhether or not the encounter took place near a streamNon-significant(X2= 0.44, d.f .=1, p= 0.51)Chi Squared testThe level of protective clothing/ footwear worn at the timeNon-significant(p=0.22)Fisher’s Exact testExtra precautions taken e.g. carrying a weapon/ care and awarenessSignificant(p<0.01)Fisher’s Exact testThe weather conditionsNon-significant(p=0.31)Fisher’s Exact testThe habitat where the encounter occurredNon-significant(p=0.97)Fisher’s Exact testThe activity the respondent was undertaking at the timeNon-significant(p=0.15)Fisher’s Exact testTable 9: Explanatory variables tested for significance in explaining the differences between those who encountered a snake and were bitten and those who encountered a snake and were not bitten.There was a significant difference between bite victims and non-bite victims as to whether or not the snake was seen first, before a potential bite incident (p<0.01), see figure 10. The majority of bite victims (91%) did not see the snake before it bit them. Figure 10: Percentage of bitten and non-bitten respondents who did/ did not see the snake before a potential bite incidentThere was also a significant difference between bite victims and non-bite victims with regards to the density of the vegetation on the ground where the respondent was walking (p<0.01) (figure 11). Most respondents who encountered the snake and were not bitten were walking through very little or no vegetation. This would increase the likelihood that the snake would be seen, as there would be less vegetation for the snake to hide in. However, 34% of bite victims were bitten in areas of little or no ground vegetation, and the majority of these people did not detect the snake. Although there was little vegetation, the ground was often carpeted with dry leaves, which B. caribbaeus is camouflaged against. Therefore, although areas with little vegetation increase the chances of a snake being seen, it is possible that it may still remain undetected as B. caribbaeus relies on crypsis as an evolutionary adaptation to avoid detection.Figure 11: The density of the ground vegetation that bitten / non-bitten respondents were walking through at the time of the snake encounter.There was also a significant difference between the two samples in terms of whether or not the respondent had tried to kill the snake before any potential bite incident (p<0.01). The majority of respondents who had attempted to kill the snake were not bitten. Most bite victims had not attempted to kill the snake before they were bitten simply because they had not seen it. The 4% of bite victims that did attempt to kill the snake before they were bitten lost sight of the snake after approaching it, resulting in a bite. It should be noted that it is not the act of attempting to kill the snake that decreases the probability of being bitten- it is the fact that those that attempted to kill the snake had seen it and where aware of its presence. Therefore this result should not be interpreted as suggesting that attempting to kill a snake will reduce the risk of a bite.There was a statistically significant difference between bite victims and non-bite victims with regards to whether or not the snake was trodden on (p<0.01). Again, this is linked to whether or not the respondent saw the snake; non-bitten respondents did not tread on the snake because they had seen it and knew where it was.There was a statistically significant difference between bite victims and non-bite victims in terms of extra precautions that were taken (p<0.01) (figure 12). It was found that more bite victims were taking extra precautions compared to non-bite victims, such as carrying a weapon or trying to be careful where they were walking. However, in this study it seems that these extra precautions did not help decrease chances of being bitten. A possible explanation is that carrying a weapon gave the bearer a false sense of security, and maybe they were therefore slightly more reckless. Only one person used a torch, and despite the fact this respondent was bitten it should not be considered that such a precaution is not useful, as the torched being used was a flame torch that provided little light while the respondent was walking through vegetation. Several respondents who were bitten at night reported that they had not seen the snake in the darkness.Figure 12: Other precautions (i.e. other than protective clothing or footwear) that respondents had put in place at the time the snake was encountered. Although the season, the time of day and the weather were not statistically significant in explaining differences between those people who were and were not bitten, it is interesting to learn that 79% of all encounters took place during the daytime, 72% during the dry season, and 85% when the weather was dry. However, this is probably reflective of when humans (as opposed to snakes) are most active, especially as St. Lucians are renowned for hating staying out in the rain! The extent of protective clothing worn by bite victims and non-bite victims was not statistically different; 68% of bite victims were wearing no protective clothing, compared to 66% of non-bite victims. For the bite victims who did not see the snake, the addition of protective clothing may have been beneficial in either preventing envenomation, or lessening its severity. Unfortunately, data is not available from this study to show this, however it often recommended that people wear long trousers and thick, high boots (preferably made of leather) to help protect against snakebite (Dr Jenny Daltry, pers. comm.) Although wearing protective clothing such as thick high boots is advised, these should not be relied upon as offering complete protection, as many bite victims were bitten in locations not protected by boots (e.g. the upper leg or hand), and in some cases the fangs of B. caribbaeus penetrated through rubber wellington boots, resulting in envenomation. In summary it would seem that the most important factors explaining differences between people who were bitten and people who encounter B. caribbaeus but are not bitten are related to whether or not the snake was seen. 4.4 Treatment preferences4.4.1 Medical versus non-medical treatmentFrom the general community survey of 169 respondents, it was found that in the event of a hypothetical B. caribbaeus bite, 92% of respondents would seek professional medical help as their first choice of treatment. Four percent would prefer to visit a local healer and use traditional medicine, and visit the hospital after this. Two percent of respondents would prefer to only visit a traditional healer/ use herbal remedies and would not seek medical help. A further 2% of respondents were unsure what action they would take. There were no statistically significant demographic variables that influenced these decisions (Table 10).Explanatory variableSignificanceStatistical test usedGenderNot significant(X2= 1.06, d.f=1, p=0.30)Chi Squared testWhether or not the respondent lives in B. caribbaeus rangeNot significant(X2= 0.02, d.f=1, p=0.89)Chi Squared testAgeNot significant(p=0.18)Fisher’s Exact testTable 10: Explanatory variables tested for significance in explaining the differences between those community survey respondents who would seek professional medical help, and those who would not.From the survey of 47 bite victim respondents, 60% of respondents sought medical help (i.e. hospital treatment) first before any other (i.e. traditional) treatment. However, 13% of bite victims (21% of those who sought medical help first) did not seek medical treatment straight away. Thirteen percent of bite victims sought traditional treatment (traditional healer/ home made herbal remedies) first, and only sought medical help thereafter. Twenty eight percent of bite victims sought only traditional treatment, either by visiting a traditional healer or making their own herbal remedy, and did not seek medical help at any point. In total, 66% of bite victims used traditional treatment methods at some point after they were bitten, regardless of what treatment they sought first. A larger proportion of bite victims favoured traditional treatment methods over professional medical treatment then would have been predicted by extrapolation of the results from the community based survey. Of the community survey respondents, 92% indicated that they would seek medical help as there first form of treatment, but in reality this figure dropped to 60% in bite victims, with 21% of this group delaying their arrival to hospital. The use of traditional treatment in St. Lucia is perfectly acceptable, so it is unlikely that respondents of the community survey gave untruthful answers. One possible explanation for the difference may be the effect of other people regarding the decision made on what treatment is sought first. Well meaning relatives may influence a bite victim to use an alternative to medical treatment. However, it is possible that differences are due to demand characteristics or a social desirability bias in the community survey, where respondents may change their answers to what would appear most favourable.Gender was significant in explaining the differences in treatment preference by bite victims (p<0.01; table 11). A significantly higher percentage of males (69%) sought medical help as their first form of treatment compared to females (13%). However, all female bite victims were bitten before the year 2000 (mostly in the 1960s and 1970s), and the decade when respondents were bitten was found to be a significant factor (p=0.03; figure 13). Before the 1970s traditional medicine was the preferred method of treatment, and no bite victim respondents bitten before this time chose medical help as their first form of treatment. However, medical treatment has since become more popular as the first form of treatment, particularly over the past two decades. Therefore a further factor affecting the difference between the treatment preferences of community survey respondents and those actually sought by bite victims is the decade, as bite victims included in the sample include those bitten in previous decades when attitudes to professional medicine were different, while the community survey patients reflected 2009 attitudes. However, even during the current decade (2000s), 25% of bite victims still did not seek medical treatment first, which is still much higher than that expected from looking at the community survey.Explanatory variableSignificanceStatistical test usedGenderSignificant( p<0.01)Fisher’s Exact testDecade of biteSignificant( p=0.03)Fisher’s Exact testAge when bittenNot significant(p=0.08)Fisher’s Exact testTable 11: Explanatory variables tested for significance in explaining the differences between those bite victims who did seek professional medical help, and those who did not.Figure 13: Percentage of bite victim respondents bitten in each decade who did/ did not seek medical help as their first form of treatment.4.4.2 First aid preferences From the community survey, 32% indicated that if they were bitten by B. caribbaeus, they would carry out at least one of the following: tie a tourniquet, incise the wound (or just above it) or attempt to suck out the venom either by mouth or with a suction kit. Figure 14 gives a breakdown of the methods respondents would use. All of these methods are harmful, and particularly unsuitable in response to a bite from B. caribbaeus as the venom has local effects. The recommended response would be to keep calm, carry out none of the above methods, and get professional medical treatment at Victoria Hospital as soon as possible.Figure 14: The proportions of community survey respondents who would carry out varying degrees of damaging first aid in the event that they were bitten by B. caribbaeusWhether or not the respondent lived within the range of B. caribbaeus was a significant predictor of whether respondents would carry out some form of damaging first aid (p=0.01; table 12). People living in areas within B. caribbaeus range are more likely to suggest practising some form of damaging first aid (43%) compared to those not living in areas within B. caribbaeus range (23%). This may be because it is locally believed that these practices are beneficial, and if a person lives within the B. caribbaeus range they are more exposed to the application of these methods and so more aware of them. Explanatory variableSignificanceStatistical test usedGenderNot significant(X2=0.12 , d.f=1, p=0.73)Chi Squared testWhether or not the respondent lives in B. caribbaeus rangeSignificant(X2=6.56 , d.f=1, p=0.01)Chi Squared testAgeSignificant(p=0.03)Fisher’s Exact testTable 12: Explanatory variables tested for significance in explaining the differences between those community survey respondents who would use some form of damaging first aid, and those who would not.Gender was not a significant predictor, but age was, with younger people (<40 years) more likely to carry out damaging first aid (p=0.03; figure 15). Figure 15: The percentage of community survey respondents from each age group who would/ would not carry out some form of damaging first aid technique in the event that they were bitten by B. caribbaeus From the survey of 47 bite victims, 68% of respondents underwent some form of damaging first aid administered either by themselves or with the aid of another person (figure 16). Figure 16: The proportions of bite victims who underwent varying degrees of damaging first aid in response to a B. caribbaeus biteIt is evident that a much larger proportion of bite victims underwent some form of damaging first aid (68%) then was indicated in the hypothetical responses to the community survey (32%). All bite victims lived in areas within the range of B. caribbaeus, and this variable was significant in explaining differences between community survey respondents who would/ would not practise damaging first aid. However, when comparing community survey respondents who lived within B. caribbaeus range with bite victims (all of whom lived within B. caribbaeus range), a higher percentage of bite victims underwent damaging first aid than would have been expected from the survey (68% compared to 43% respectively). This may be due to the influence of others, as for some bite victims that underwent damaging first aid, the treatment was administered by another person, or it may be due to issues with biased responses to the survey (i.e. social desirability bias or other founder effects). However, in this case, it may have been less obvious to the respondents what answers were expected. None of the explanatory variables tested were significant in explaining the difference between people who underwent some form of damaging first aid, and those who did not (table 13). Factors affecting this difference amongst bite victims are therefore unclear.Explanatory variableSignificanceStatistical test usedGenderNot significant( p=0.70)Fisher’s Exact testDecade of biteNot significant( p=0.14)Fisher’s Exact testAge when bittenNot significant(p=0.68)Fisher’s Exact testTable 13: Explanatory variables tested for significance in explaining the differences between those bite victims who used some form of damaging first aid, and those who did not.4.5 Attitudes and behaviour towards B. caribbaeus 4.5.1 Local attitudes to B. caribbaeusEighty one percent of community survey respondents reported that they strongly disliked or disliked B. caribbaeus, and only 4% liked it (figure 17). This highlights just how unpopular B. caribbaeus is in St. Lucia. Figure 17: Respondent attitudes towards B. caribbaeus regarding how much they like/ dislike the species The main reasons given for such negative attitudes were that B. caribbaeus is poisonous, it bites and is therefore dangerous, that the respondent is very scared of it, that (in the respondents opinion) the species will kill you if it bites, and that it is a creature cursed by god and one that tricked Eve in the Garden of Eden.Respondents living within the species range more strongly disliked B. caribbaeus than those living outside of the species range (figure 18; table 14). Overall, 87% of respondents from within B. caribbaeus range held a negative view of the species (strongly disliked + disliked), compared to 75% of respondents living outside of the range. A higher percentage of respondents living outside of the species range felt neutral about B. caribbaeus or were unsure of their feelings towards it. The main explanation for this was that these respondents did not have contact with the species; therefore either did not know much about it or reported that it had never bothered them. Interestingly, 6% of respondents living within B. caribbaeus range reported that they liked the species, compared to 3% of respondents living outside of these areas.Figure 18: Attitude towards B. caribbaeus by respondents living in inside/ outside of the species rangeExplanatory variableSignificanceStatistical test usedGenderNot significant( p=0.82)Fisher’s Exact testWhether or not the respondent lives in B. caribbaeus rangeSignificant( p<0.01)Fisher’s Exact testAgeNot Significant(p=0.45)Fisher’s Exact testTable 14: Explanatory variables tested for significance in explaining the differences between community survey respondents with different opinions of the species4.5.2 Local attitudes towards hypothetical conservation protection Interestingly 31% of respondents felt positively towards (hypothetical) potential plans to protect B. caribbaeus, but only if it meant that humans and snakes could somehow be separated so that the two did not come into contact. Protection of B. caribbaeus without attempts to keep people and snakes apart was generally not accepted, as the species is considered a danger to human health and many of these respondents said that B. caribbaeus would not be tolerated if found in or around settlements. Fifty five percent of respondents remarked that they would not be supportive of any attempt to protect B. caribbaeus, regardless of how it was carried out. Finally, 14% of respondents were unsure how they felt about any prospect of protection. None of the tested explanatory variables were statistically significant in explaining differences between those would not be supportive of any form of protection and those that did not have this opinion (table 15).Explanatory variableSignificanceStatistical test usedGenderNot significant(X2=0.61, d.f.=1, p=0.43)Chi Squared testWhether or not the respondent lives in B. caribbaeus rangeNon significant( X2= 2.57, d.f.= 1, p=0.11)Chi Squared testAgeNot Significant(p=0.09)Fisher’s Exact testTable 15: Explanatory variables tested for significance in explaining the differences between community survey respondents who would and would not be supportive of conservation protection.4.5.3 Abilities in identifying B. caribbaeusOut of the 169 community survey respondents, 58% claimed that that they were able to correctly identify B. caribbaeus, and could tell the difference between this species and the Boa Constrictor (Boa constrictor orophias) which also inhabits St. Lucia and has cryptic colouration. However, 42% admitted that they could not tell the difference between the two species. This finding is important, as the Boa Constrictor is a protected species in St. Lucia. As many people kill snakes on sight, for fear that it may be B. caribbaeus, the finding that over 40% of sampled respondents can not tell the difference between the two species may be a conservation problem. However, judgement on a respondent’s ability to identify the species has been based on the respondent’s word. This is due to the limitations associated with using methods such as picture cards, as respondents may chose pictures based on (for example) the positioning of the animal (Dr. D Warrell, pers. Comm.). Respondents were asked which physical/ behavioural features they considered when identifying the species; however the answers have not been used to judge the respondents ability at identification, as the researchers judgement on this may be more inaccurate than if the respondent’s word is taken.Seventy two percent of respondents from areas within B. caribbaeus range report that they are able to identify the species, compared to 49% of those living outside.Sixty nine percent of males claimed that they were able to identify the species compared to 44% of females. Age was not a significant predictor of reported identification ability (table 16).Explanatory variableSignificanceStatistical test usedGenderSignificant(X2=10.83, d.f.=1, p<0.01)Chi Squared testWhether or not the respondent lives in B. caribbaeus rangeSignificant( X2=8.49 , d.f.= 1, p<0.01)Chi Squared testAgeNot Significant(p=0.19)Fisher’s Exact testTable 16: Explanatory variables tested for significance in explaining the differences between community survey respondents who could and could not identify B. caribbaeus.4.5.4 Respondent behaviour upon encountering B. caribbaeusThirty three percent of respondents stated that they would attempt to kill B. caribbaeus if they encountered one. Age was a significant factor affecting this (p<0.01; table 17) with a significantly higher percentage of males reporting that they would attempt to kill the snake (46%), compared with only 16% of females. Age was also a significant factor. Figure 19 shows that respondents over 60 years old felt the most inclined to kill B. caribbaeus on sight (61%), followed by respondents under the age of 17 (50%).Explanatory variableSignificanceStatistical test usedGenderSignificant(X2=16.81, d.f.=1, p<0.01)Chi Squared testWhether or not the respondent lives in B. caribbaeus rangeNot Significant( X2=3.04 , d.f.= 1, p=0.08)Chi Squared testAge Significant(p<0.01)Fisher’s Exact testTable 17: Explanatory variables tested for significance in explaining the differences between community survey respondents who would and would not kill B. caribbaeus.Figure 19: Percentage of respondents in each age group who would/ would not kill B. caribbaeus if encountered.For the purpose of comparison, these attitudes were compared with the outcomes of actual reported encounters with B. caribbaeus. Of these reported encounters, 32% of respondents did actually kill the snake, 17% reported that another person was called to kill it, and 51% reported that the snake (as far as they were aware) was not killed. The percentage of respondents that did kill the snake is remarkably similar to the percentage of those who stated that they would intend to kill it if it was encountered. Only around half of snakes encountered survived, and even this is possibly an over estimate as any B. caribbaeus encountered around a community would almost certainly be killed by someone. None of the explanatory variables tested where significant factors in explaining any differences between those respondents who did kill B. caribbaeus when encountered, and those who did not (table 18).Explanatory variableSignificanceStatistical test usedGenderNot Significant(p=0.27)Fisher’s Exact testWhether or not the respondent lives in B. caribbaeus rangeNot Significant( p=0.75)Fisher’s Exact testAge Not Significant(p=0.62)Fisher’s Exact testTable 18: Explanatory variables tested for significance in explaining the differences between respondents who did and did not kill B. caribbaeus.4.6 Examining geographic variation in human- B. caribbaeus conflictFigures 20 and 21 reflect the frequencies and the decades of sightings reported by respondents of the community survey, while figure 22 reflects sightings reported by wildlife professionals. Figures 23 and 24 reflect the frequencies and decades of bite incidents reported by bite victims from the snowball sample, and finally figure 25 combines the locations from figures 20-24 in order to produce a map to reflect B. caribbaeus estimated current range.The estimated range includes areas around Grande Anse and Louvet on the Northeast coast (see figure 26 for location names), as well as a belt across the island from Dennery on the East coast across to Canaries and Anse la Raye on the West coast, encompassing areas in the middle of the island such as Millet, but not including areas of very high elevation such as Mount Gimie. From figure 25 it can be seen that the mapped locations span both dry, mesic and rainforest habitats. However, this habitat map is fairly crude and does not reflect the fact that B. caribbaeus are encountered in a large variety of habitats, such as in plantations, in gardens and even in settlement areas, as well as in the forested areas shown on the parison between figures 25 and 27 shows a large retraction in species range, with a previously almost island-wide distribution now being limited to a strip across the middle third of the island as well as a patch around the Northeast coast.Conflict between humans and B.caribbaeus could potentially occur anywhere within the estimated current range. However, figures 23 and 24 show that hotspot areas for snakebite seem to be clustered around areas such as Canaries, Anse Galet and Anse la Raye on the West coast, around Millet and the Roseau dam in the centre, around areas surrounding Dennery and Praslin on the East coast such as Bordelaise and St. Joseph estate as well as at a couple of locations on the Louvet estate and close to Grande Anse on the Northeast coast. Figure 23 shows that the location with the highest frequency of bites was on the West coast in the hills surrounding Canaries, near a place known as “Esperance”. A total of 5 bites were recorded here within the same 1 km2 grid cell. Figure 24 shows that the majority of recent bites are occurring around Canaries and Anse la Raye on the West coast, around Bordelaise and St. Joseph estate on the East coast, and in Millet in the centre. Bites close to Grande Anse in the Northeast have not been reported for at least a decade, and bites within the village of Praslin (just south of Dennery and Bordelaise) have not occurred for at least 30 years.In the community of Praslin, local respondents unanimously stated that B. caribbaeus had declined around the community, and is no longer encountered within the village. Some locals suggested that the decline has increased recently as a result of the development of a large holiday resort in the Praslin Bay area (Westin Le Paradis Resort). A local environmental scientist commented that the area “needed to become Fer-de-Lance free” and that by June 2009 at least 118 B. caribbaeus had been killed on this development site.Figure 20: Map to show the frequency of B. caribbaeus sightings reported by respondents in the general community study.Figure 21: Map to show the most recent decade of sightings made at each location, as reported in the general community survey Figure 22: Map to show sightings of B. caribbaeus by wildlife professionals within the last ten yearsFigure 23: Map to show the locations and frequencies of snake bite incidence as reported by bite victims from the snowball survey Figure 24: Map to show the decade of the most recent bite incident that occurred at each location, as reported by bite victims from the snowball sampleFigure 25: Map to show all sightings and bite incidents mapped against forest habitat. Forest habitat GIS layer courtesy of Matthew Morton, Durrell Wildlife Conservation TrustFigure 26: Map of St. Lucia identifying location namesCourtesy of Matthew Morton4.7 Investigating historical evidence of past species rangeFigure 27 maps the inferred historical range of B. caribbaeus. Each location on the map was identified from articles from historical publications (as detailed in the methods section). This figure shows that the species was once probably found almost island wide, maybe excluding the very dry northern and southern tips of the island, as well as areas of high elevation.The lack of points in the interior of the island is probably a reflection of the fact these areas are dense rainforest, which is a place few people ventured into (see habitat map, figure 25). For this reason, there are likely to have been fewer human-snake encounters in these regions, so the lack of documented B. caribbaeus occurrence here should not be interpreted as absence of the species. Even in the 1930s, the species appears to be found island wide. Despite this, there is documented evidence that B. caribbaeus was decreasing in abundance, and was even considered a rarity. As species abundance is not reflected in the distribution map, a selection of examples suggesting declining abundance has been presented below:YearQuotation from historical publicationPublication1897“Two persons are reported as having died from the bite of the Fer-de-lance, now becoming rare in the Island.”Annual Report on the St. Lucia Blue Books1900“The Morne and Vigie where our garrison is now barracked, once infested with snakes, have been so effectively cleared that not a single one has ever been seen”.The Voice of St. Lucia1901“There is almost a total disappearance of Fer-de-lance from all other districts on the island” (apart from Soufriere district).The Voice of St. Lucia1906(The fer-de-lance has) “practically disappeared from every part of the island which has been cleared of cover for them to lurk in.” The Voice of St. Lucia1906“…the valley was so infested with snakes that certain cane pieces had to be fired before cutters could go into them. For several years now it has not been heard of a single Fer-de-lance being found in any cultivated part of the valley.”The Voice of St. Lucia1911“A few years after the introduction of the mongoose, the number of snakes rapidly diminished, until none but a few large ones were to be found.”St. Lucia Gazette1911(The Fer-de-lance) “for some years had practically ceased to exist anywhere but in the heavy, practically unexplored forest lands.”St. Lucia Gazette1923“The most dangerous snake, the Fer-de-lance, is being exterminated through the introduction of the Mongoose and the advance of cultivation. In addition, the Government pays 6d. for every serpent head brought to the various police stations, and this helps considerably in the process of extirpation.”St. Lucia Gazette1924“With the destruction of vipers by the mongoose aided by wholesale slaughter by the inhabitants, who are stimulated by the Government reward of sixpence for every vipers head brought to the police, the Fer-de-lance has almost totally disappeared from St. Lucia.” St. Lucia GazetteFigure 27: Map to show the expected historical range of B. caribbaeus as identified by locations given in historical publications. Squares are colour-coded to reflect the decade of the last sighting.5. DiscussionAlthough snake bite in St. Lucia is a fairly uncommon event, it is still a threat to human health and welfare, and the ensuing conflict between people and B. caribbaeus is also detrimental to the conservation of this snake species. In order to protect people and ensure the preservation of the species, the current conflict levels must be reduced.5.1 Decreasing snakebite riskBeing unaware of the presence of B. caribbaeus proved to be the most significant factor in increasing bite risk. It is important to minimise this risk by increasing the chances of spotting the snake first. However, people who saw snakes often killed them, so before snakebite risk mitigation measures can be put in place, it is imperative that education and awareness raising programmes are implemented to change these habits (see section 5.2).Results suggest that people should avoid walking through vegetation, especially thick secondary growth as this will decrease the chance that a snake will be seen. Even in areas devoid of vegetation it should be recognised that B. caribbaeus camouflages perfectly against dry leaves, so great care must still be taken. The species will happily hide in crevices and under objects such as rocks, logs, coconut husks and large Coccoloba pubescens (“Gran fèy”) leaves, so people should be very wary when stepping amongst such obstacles. Hands and feet should not be placed where they can’t be seen. It should also be known that B. caribbaeus will climb, so people must be aware that the species may be found in the trees. It is advised that people walk on wide, clear paths/ trails and avoid walking close to any vegetation at the sides. B. caribbaeus can easily bite if it strikes from the side of a narrow path with densely vegetated edges. If going out at night, even if only stepping out of the house a powerful torch should be used in order to see what is on the ground and avoid any potential snakes.The majority of bite victims were bitten on the foot or lower leg, and most of these people were not wearing any protective clothing or footwear. It is advised that some form of protection is worn such as high thick leather boots and long, loose trousers, as in the event that B. caribbaeus is not seen first these may help protect against snakebite. However it should be noted that the fangs of B. caribbaeus may penetrate rubber boots, and that other areas of the body such as the upper leg or hands have no protection. If “snake chaps” or “snake gaiters” are available, these should be worn. The possibility of somehow making protective clothing such as snake chaps available to the public should be considered, as currently only SLFD range workers wear them. Although such equipment is expensive and would need to be imported, the cost of snake chaps will be much cheaper than the cost of antivenom! If gardening, thick gardening gloves should be worn as they may offer some protection, although they may still be penetrated in the event of a bite. 5.2 Addressing treatment preferences, attitudes and behavioursA large proportion of bite victims did not immediately seek medical help, and many practised damaging first aid such as tying a tourniquet. Techniques such as tourniquets and suction have been widely warned against in the medical literature (e.g. Theakston, 1997; Holstege and Singletary, 2006). It is advised that bite victims keep calm, carry out no such damaging first aid, and travel to Victoria hospital (the only hospital that treats snake bite) as soon as possible. To address these treatment preferences it is necessary to implement awareness and education programmes. This could be combined with a B. caribbaeus conservation education programme, as before any physical conservation measures can be implemented, it is essential to increase knowledge and positive attitudes towards B. caribbaeus. Currently, the majority of local people feel negatively about the species, and many respond to an encounter by killing the snake. Without changes in attitudes, it is unlikely that conservation efforts will be supported by the public, and therefore are unlikely to succeed. However, as time and resources would have to be spent implementing such a programme, it is essential that the methods used are effective and successful. Morgan and Gramann (1989) carried out a study to investigate the effectiveness of a snake oriented wildlife education programme on students in Pennsylvania, United States of America. It was found that “mere exposure” (where the object (in this case a snake) is simply kept in a tank and allowed to be viewed by the target audience), combined with “modelling” (which involves the audience observing a demonstrator interacting with the animal) were the most effective methods at significantly improving attitudes towards snakes. However, Morgan and Gramann (1986) suggest a “full treatment approach” when implementing an education programme, which should also include an opportunity for the target audience to physically interact with the species, as well as presentation of factual material.Based on the results found by Morgan and Gramann (1989), when implementing an education programme related to B. caribbaeus it may be beneficial to use a combination of methods. Factual information about the species, and advice on how to prevent snake bite and what to do if bitten could be presented using video media, or as a slide show and in leaflets and posters. The audience could be given an opportunity to safely observe B. caribbaeus that are kept in large, secure glass vivariums. As the species is venomous, it would not be suitable for handling. However, the native Boa Constrictor is a fairly docile and harmless species, and this species could be used for educational handling purposes as a substitute for B. caribbaeus in order to increase positive attitudes towards snakes. This could also raise the profile of the protected Boa Constrictor, and aid in conservation efforts for this species as well as for B. caribbaeus. Currently there seems to be no formal education programmes for these species, although a captive Boa Constrictor is kept at St. Lucia’s mini-zoo, which is visited by school children. Staff involved in the implementation of the programme would need to themselves have positive attitudes towards B. caribbaeus, and it may be beneficial for a popular public figure to be involved with the programme (Morgan and Gramann, 1986). The programme would need to be carefully and thoroughly planned and implemented, and the outcomes and success of the education programme should be monitored in order to assess whether or not it increases positive attitudes and behaviour towards snakes. 5.3 Distribution mapsThe inferred current range and the highlighted conflict areas reflect locations where conflict mitigation management strategies are most needed. People need to be made aware of B. caribbaeus distribution, so that when travelling in these areas precautions can be taken to minimise the risk of snake bite. The inferred current range also demonstrates the decline in species range over time, which highlights the urgency for conflict to be reduced and for conservation measures to be implemented.5.4 Other potential conflict mitigation methods5.4.1 TranslocationIn order to reduce human-snake conflict, a number of conservation projects involving various Rattlesnake species have been implemented in America and Canada. The methods usually involved the translocation of problem and vulnerable snakes. For example, as part of the Lethbridge Rattlesnake Conservation Project in Canada, Ernst (2004) documented how Prairie Rattlesnakes (Crotalus viridus viridus) were removed from urban areas in order to avoid potential conflict situations, and relocated to Cottonwood park, where they were implanted with Passive Integrated Transponder (PIT) tags for long-term identification and monitoring and underwent a soft release. A drift fence with snake traps was erected at locations along the park boundaries to capture any snakes migrating towards residential areas. The project at the time of the published report was still in its infancy, and despite some encouraging results the successfulness of the scheme was still undetermined. Brown, Bishop and Brook (2009) investigated the effectiveness of short-distance translocation (SDT) of problem and vulnerable Western Rattlesnakes (Crotalus oreganus) in British Columbia, Canada. Short distance translocation was found to be an ineffective tool as a long-term solution to reducing human-rattlesnake conflict, as within a short period of time most of the snakes that underwent SDT returned to the area that they were caught from, a finding also reported by Hardy Sr. et al (2001). However, it was found that even multiple SDT’s did not seem to affect behaviour, body condition or mortality rates, unlike long distance translocations (LDT’s) involving the transport of the animal to a place outside of its home range (Brown, Bishop and Brook, 2009). The United States Geological Survey (USGS, 2006) report that relocation of rattlesnakes has been an ineffective method of removing problem animals, and a site with a high abundance of prey species will just attract other snakes to the area. They have therefore suggested that the best methods for reducing potential conflict is to make areas of human activity unattractive to snakes and prey species such as rodents. LDT has been used in St. Lucia as a method of removing “nuisance” Boa constrictors (M. Morton, pers. comm.). However, LDT of snakes to a location outside of their home range is not recommended as a method for reducing human-snake conflict, due to effects on life history, spatial ecology and mortality rates (Nowak et al, 2002; Butler et al, 2005). Snakes that undergo LDT are likely to move away from their release site, and travel large distances, which is likely to bring them back into contact with humans. The chances of survival for these individuals will be lowered by LDT. Short distance translocation of the animal to a location within its home range and away from human activity has much less of a detrimental effect, compared to LDT’s. However, SDT’s will not be a long-term solution, as the individual is likely to return. Despite this, non-removal of a venomous snake from a potential conflict area (e.g. close to inhabited areas) is also not a desirable option, as the risk of human-induced snake mortality is high and there is also a risk to human health. For this reason, SDT of the animal into good quality habitat less than 500m from where it was collected may be a potential short-term solution for removing B. caribbaeus away from any immediate risk of mortality. However, as suggested by Brown, Bishop and Brooks (2009), this technique should be seen as a last resort, and only a short-term, temporary solution. If SDT is chosen as a management option, a response team would need to be available to quickly respond to reports of problem snakes, and the initiative would need to have the full support of locals, which could be obtained either through education and awareness programmes or some form of incentive scheme. Post-release monitoring is an essential part of the translocation process, as data is vital for assessing success, and guiding future improvement. This may be implemented by implanting PIT tags or other forms of radiotransmitters which can allow for individual identification and monitoring, and would help increase knowledge on species ecology, for example range size and habitat use. It would also be useful to implement experimental testing of the practises used, and for results to be reported (Germano and Bishop, 2008). An experimental pilot study using the methods above could for example, start off around a particular community (e.g. Millet, where there is high levels of conflict around the community, plus a SLFD outpost where staff would be available for quick response). 5.4.2 Humane snake trapsIn order to reduce conflict in and around homes, re-useable and humane snake traps should be distributed and placed along walls inside/ around settlements. This may be beneficial for intercepting snakes before conflict occurs. Snakes will often move along a vertical object, for example a solid fence or a wall, and will seek shelter in the specially designed trap which is placed in its path. The snake could then be safely removed from the home, and released a short distance away into suitable habitat, as suggested above. If humane snake traps were distributed to local communities, training would be needed to ensure safe and humane use. 5.4.3 Barrier methodsTo prevent the movement of the brown tree snake (Bioga irregularis) into sensitive areas on the island of Guam (where it was accidently introduced), several effective snake barriers have been designed to control dispersal into such locations (e.g. Perry et al, 1998; Rodda et al, 2007). The implementation of carefully designed and tested snake barriers in St. Lucia may be another possible method in reducing conflict, as they can be used to protect certain areas from snakes. However, the barriers were designed in order to limit the spread of a damaging invasive species (Bioga irregularis) in Guam, and the use of such barriers in St. Lucia would not be beneficial for promoting pride in the endemic snake species, as shutting them out reinforces the idea that the species should be feared. It would however, help provide separation of people and snakes, and could maybe be built around homes and schools if necessary. It would not be practical to build such structures to surround entire communities, as for the barriers to be effective there must be no gaps, and any human access points have to be carefully designed and cautiously used. They would therefore only be suitable on a very small scale. Drift fencing can also be used as another barrier method to intercept snakes before they enter an area where they are unwanted.5.4.4 Deterring snakes from community areasIt would be useful to make community areas unattractive to snakes and their prey, and therefore decrease the chance of B. caribbaeus being found in these locations were conflict potential would be high.Vegetation around communities and in “gardens” (“gardens” in St. Lucia are private plots of land equivalent to British allotments, where fruits and vegetables such as bananas and mangos are grown) should be managed and kept tidy in order to deter B. caribbaeus from hiding in it, or to increase the chance of detecting it. Piles of wood or debris around “gardens” or communities should be cleared to limit potential hiding places for snakes. B. caribbaeus is probably attracted to areas where rodents are abundant, as rats and mice seem to be a favourite prey species. Therefore it is advisable that efforts are made not to attract rodents, by carefully disposing of any food/ edible waste and by practising pest control.5.4.5 Respecting B. caribbaeus in its natural habitatSnakes are currently not tolerated in St. Lucian communities, where they are considered a threat to society. However, in natural habitat within the protected rainforest reserve in the heart of the island, snakes are frequently killed on sight, actions that are supported even by government agencies. As the maps compiled in this study show, there has been a decrease in the range of B. caribbaeus. The protected rainforest reserve could potentially become one of the last natural strongholds for the species, and as it is an area away from high population density, it is advised that B. caribbaeus in these areas are at least respected. The SLFD moto “take nothing but photographs, leave nothing but footprints, kill nothing but time”, which is displayed on forestry reserve signs should be adhered to in these protected areas, by all parties involved. In order to avoid possible conflict within natural habitat, it may be useful to display educational signs that increase conservation awareness, as well as detailing useful precautions that can be taken to avoid snakebite, such as wearing protective clothing and keeping to trails. 5.5 Limitations5.5.1 General community surveyOne of the main limitations when carrying out the community survey was the use of opportunistic sampling to recruit respondents. This method potentially introduces bias, for example because the sample of people was based on who was available at the time of the study (i.e. during working hours on week days). Also, the area of the community where the study took place may also introduce bias, as only certain types of people may congregate there. It is also possible that only people who seemed approachable were sampled using this method.Other potential problems include researcher bias, where the respondent is (often unintentionally) influenced by the person carrying out the investigation, as well as possible demand characteristics or a social desirability bias, where respondents may change their answers to what would appear most favourable. For example, a respondent may say that they would seek medical help if bitten by a snake in order to conform with perceived expectations, even though in reality they would not choose this treatment. Different researchers may have different effects on respondents, and as there were two people carrying out the surveys (myself and a local field assistant), this may have created further bias. Between seven and twelve respondents were interviewed in each community, with the number of people sampled reflecting how busy or quiet a community was. This difference in number of people sampled at each location will introduce further bias.The main logistical limitations associated with the community survey involved time limitations, for example only a limited number of hours could be spent at each settlement.However, despite the limitations, this community level research was able to produce a large amount of information regarding B. caribbaeus sightings as well as respondent attitudes and behaviours. 5.5.2 Bite victimsLimitations associated with the snowball sampling methods used include the fact that the sample is not representative of the wider population, and that the sampling frame cannot be identified. Also, bite victims who are well know around the community (maybe due to personality traits) are more likely to be recruited than bite victims who are less well know, maybe because they live in a secluded location or have a quiet or unsociable disposition. However, despite this, the use of a community based study allowed for a greater amount of data to be obtained than if hospital records alone had been sampled, and the use of the snowball survey method allowed this hard to reach subpopulation to be surveyed.5.5.3 Hospital recordsThere were logistical limitations involved when searching hospital records, and only 35 hours could be dedicated to this sampling. Records were searched as systematically as possible, however this was occasionally affected by the disorganisation of these files. Some information, such as the outcome of snakebite was often not recorded, which made analysis more difficult.5.5.4 Distribution maps The distribution maps can only infer the current and historic range of B. caribbaeus. The mapped locations may reflect the degree of human activity within an area, which is evident from the map of historical range as there are very few locations reported for rarely visited rainforest areas. However, to minimise this problem when mapping current distribution, wildlife professionals who work in these remote areas were sampled.Locations are mapped as 1km2 squares on the maps, which is fairly coarse grain. This was due to difficulties in determining the exact point locations from historical records, as well as for sightings identified from the general community survey, as respondents often had difficulties interpreting the map. Sightings and bite locations are based purely on respondent recall, and rely on accuracy here. Also, the map of historical distribution relies on the accuracy of locations reported in the historical publications. However, despite this, the current range maps are useful in identifying potential conflict areas that must be targeted with conflict mitigation management strategies, and through mapping of historical data it has been possible to present evidence on a decline in range, and therefore a possible conservation need.5.6 Further researchIt is important that the reasons behind B. caribbaeus decline are investigated, in order to reverse this trend. Potential factors for investigation include human persecution, human development and associated decrease in suitable habitat, the impacts of introduced mongoose on B. caribbaeus (particularly effects on juvenile snakes which are potentially more vulnerable), the effects of agricultural chemicals, and the possible effects of the introduced venomous cane toad (as B. caribbaeus may eat amphibians, and Cane toads have been found to cause declines in frog-eating snakes (Curry, 2000)). The conservation status of B. caribbaeus needs to be investigated and determined in order to formally quantify any potential conservation need. Research must be carried out to investigate the species ecology, as there is currently very little information on this, and the knowledge gained will be important in understanding reasons for decline and help to guide conservation efforts.Conservation measures must be implemented (such as conservation education programmes), and it would be useful if experimental designs are used in order to assess successfulness and guide future improvements.This investigation was unable to study mortality rates associated with B. caribbaeus bites, and further research could be carried out to quantify death rates and identify risk factors that increase the risk of this outcome.5.7 Conclusions Human- snake conflict on the island of St. Lucia is a concern to both human health and the conservation of B. caribbaeus. It has been found in this study that the species is widely disliked, and often killed on sight. However, snake bite is a traumatising event and can be costly to the victim. It was found that to reduce the risk of snakebite, one of the most important factors is to ensure that you see the snake first. However, the problem is that when humans do see snakes, the probable outcome is human-induced snake mortality. It has been found that the range of B. caribbaeus has decreased over time, which highlights the need for conservation effort. The current conflict levels must be addressed, and due to present attitudes on the species, it is important that this starts with education and awareness raising on issues regarding B. caribbaeus conservation, bite prevention and bite treatment. From this platform, it may then be possible to implement further conservation measures, with the support of local people. ReferencesAvico, U., Kaplan, C., Korczak, D. and Van Meter, K. (1988) Cocaine epidemiology in three European Community cities: a pilot study using a snowball sampling methodology, Brussels: European Communities Health DirectorateBrooker, S., Hay, S.I., Bundy, D.A.P. (2002) Tools from ecology: useful for evaluating infection risk models? Trends in Parasitology. 18: 70-74Brown, J.R., Bishop, C.A., Brooks, R.J. (2009) Effectiveness of Short-Distance Translocation and its Effects on Western Rattlesnakes. Journal of Wildlife Management 73 (3): 419- 425Campbell, J. A., Lamar, W. W. (2004) The Venomous Reptiles of the Western Hemisphere. Comstock Publishing Associates: Ithaca, NY.Cheng, A.C., Winkel, K.D. (2001) Snakebite and antivenoms in the Asia Pacific: wokabaut wantaim, raka hebou (“walking together”). Med J Aust 175: 648-651Curry, B.J. (2000) Snakebite in tropical Australia, Papua New Guinea and Irian Jaya. Emergency Medicine. 12: 285-294Distefano, E. (2005) Human-wildlife conflict worldwide: collection of case studies, analysis of management practises and good practices. Rome: SARD Initiative Report., D. (2008) Seeking Serpents: A mark-recapture study of brown snakes (Pseudonaja textilis) in an urban reserve. Presentation at National Wildlife Rehabilitation Conference, Canberra, 21-25 July 2008. Available from: , R.D. (2004) Lethbridge rattlesnake conservation project: 2003 progress report. Alberta Sustainable Resource Development, Fish and Wildlife Division, Alberta Species at Risk Report No. 81, Edmonton, AB.Germano, J.M., Bishop, P.J. (2008) Suitability of Amphibians and Reptiles for Translocation. Conservation Biology, 23 (1): 7-15Hay, S.I., Snow, R.W. (2006) The Malaria Atlas Project: Developing global maps of malaria risk. PLoS Med. 3 (12) e473Hardy Sr., D.L., Greene, H.W., Tomberlin, B., Webster, M. (2001) Relocation of Nuisance Rattlesnakes: Problems Using Short- Distance Translocation in a Small Rural Community. The Newsletter of the Colorado Herpetological Society. 28 (9) , A.K., Mandal, M., De, M.K., Mukherjee, H., Hati, R.N. (1992) Epidemiology of snake bite in the District of Burdwan, West Bengal. J Indian Med Assoc 90: 145-147Holstege, C.P., Singletary, E.M. (2006) Images in emergency medicine. Skin damage following application of suction device for snakebite. Annals of emergency medicine 48 (1): 105-113Kasturiratne, A., Wickremasinghe, A.R., de Silva, N., Gunawardena, N.K., Pathmeswaran, A., et al. (2008) Estimating the global burden of snakebite: A literature analysis. PLoS Med. 5: e218Lalloo, D.G, Trevett, A.J., Seweri, A., Naraqi, S., Theakston, R.D.G., et al (1995) The epidemiology of snake bite in Central Province and National Capital District, Papua New Ginea. Trans R Soc Trop Med Hyg 89; 178-182Lazell, J.D. (1964) The Lesser Antillian Representitives of Bothrops and Constrictor. Bull. Mus. Comp. Zool., Harvard Univ., 132 (3): 245-273 Long, E.G. (1974) The Serpent’s Tail: Reptiles and Amphibians of St. Lucia. University of West Indies, Iouanola series. Marques, O.A.V., Martins, M., Sazima, I. (2002) A new insular species of pitviper from Brazil, with comments on evolutionary biology and conservation of the Bothrops jararaca group (Serpentes, Viperidae). Herpetologica, 58(3): 303-312 Martins, M., Marques, O.A.V., Sizima, I. (2001) Ecological and phylogenetic correlates of feeding habits in Neotropical pitvipers of the genus Bothrops. In: Shuett, G.W., H?ggren, M., Douglas, M.E., Greene, H.W. (Eds.), Biology of the Vipers. Eagle Mountain Publishing. Eagle Mountain, UT: 307-328.McNamara, R. P. (1994) The Times Square Hustler: Male Prostitution in New York City, Westport: Praeger.Molesworth, A.M., Harrison, R., David, R., Theakston, G., Lalloo, D.G. (2003) Geographic Information System mapping of snakebite incidence in northern Ghana and Nigeria using environmental indicators: a preliminary study. Trans R Soc Trop Med Hyg.97: 188-192Morgan, M.J., Gramann, J.H. (1989) Predicting effectiveness of wildlife education programmes: A study of student’s attitudes and knowledge towards snakes. Wildl. Soc. Bull. 17: 501-509Nowak, E.M., Hare, T., McNally, J. (2002) Management of “nuisance” vipers: effects of translocation on western diamond-backed rattlesnakes (Crotalus atrox). In: Campbell, J.A., Brodie, E.D. (eds.) Biology of the vipers. Eagle Mountain, Eagle Mountain, Utah, USA.Perry, G., Campbell, E.W., Rodda G.H., Fritts, T.H. (1998) Managing island biotas: Brown Treesnake control using barrier technology. Pp. 138-143 In Proceedings of the 18th Vertebrate Pest Conference. Baker, R.O. and A.C. Crabb (Eds.). University of California, Davis, California, USA.Pollok, M., Schlitz, M. A. (1988) Does voluntary testing matter? How it influences homosexual safer sex. Paper presented at the Fourth International Conference on AIDs, Stockholm, Sweden.Rodda, G.H., Farley, J.L., Bischof, R., Reed, R.N. (2007) New developments in snake barrier technology: Fly-ash covered wall offers a feasible alternative for permanent barriers to Brown Treesnakes (Bioga irregularis). Herpetological Conservation and Biology 2 (2): 157-163Salganik, M.J., Heckathorn, D.D. (2004) Making Unbiased Estimates from Hidden Populations Using Respondent-Driven Sampling. Sociol. Methodol. 34 (1): 193-240Sasa, M., Wasko, D.K., Lamar, W.W. (in press) Natural history of the terciopelo Bothrops asper (Serpentes: Viperidae) in Costa Rica. Toxicon. Schwartz, A., Henderson, R.W. (1991) Amphibians and Reptiles of the West Indies. Descriptions, Distributions and Natural History. Gainsville, University of Florida Press. Theakston, R.D. (2007) An objective approach to antivenom therapy and assessment of first-aid measures in snake bite. Ann. Trop. Med. Parasitol. 91 (7): 857-65 Theakston, R., Warrell, D. (2000) Crisis in snake antivenom supply for Africa. The Lancet. 356: 2104U.S. Geological Survey fact sheet (2006) Reducing Rattlesnake-Human Conflicts. Available at , S.K., Chappuis, F., Jha, N., Bovier, P.A., Loutan, L., Koirala, S. (2004) Impact of snake bites and determinants of fatal outcomes in Southeastern Nepal. Am. J. Trop. Med. Hyg. 71 (2): 234-238Warrell, D.A., Arnett, C. (1976). The importance of bites by the saw-scaled carpet viper (Echis carinatus): epidemiological studies in Nigeria and a review of the world literature. Acta Tropica. 33: 307-341Watt, G., Padre, L., Tuazon, L., Hayes, C.G. (1987) Bites by the Philippine cobra (Naja naja philippinensis): an important cause of death among rice farmers. Am J Trop Med Hyg 37: 636-639Whitaker, P.B., Shine, R. (2000) Sources of mortality of Large Elapid Snakes in an Agricultural Landscape. Journal of Herpetology 34 (1): 121-128Woodroffe, R., Thirgood, S., Rabinowitz, A. (2005) The impact of human-wildlife conflict on natural systems. In: Woodroffe, R., Thirgood, S., Rabinowitz, A. (eds.) People and Wildlife: Conflict or Coexistence? Cambridge University Press.Wüster, W., Thorpe, R.S., Salom?o, M.G., Thomas, L., Puorto, G., Theakston, R.D.G, Warrell, D.A. (2002) Origin and phylogenetic position of the Lesser Antillian species of Bothrops (Serpentes, Viperidae): biogeographical and medical implications. Bull Nat His Mus Lond (Zool) 68: 101:106WWF report (2008) Common Ground: Solutions for reducing the human, economic and conservation costs of human wildlife conflict. Available at 1: General community survey questionnaireCommunity questionnaire 2009Name of person conducting interview ……………………………………………….Today’s date …………………………………………………………………………………….Location of interview ………………………………………………………………........GPS …………………………………………………………………………………………………..The person’s age: ? Child ? Adult up to 40 ? Adult between 40-60 ? Adult over 60Gender: ? Male ? Female Have they ever seen a Fer-de-Lance? ? Yes ? No ? Don’t knowIf yes- Where was the location? ....................................................................................GPS ……………………………………………………………………………………………………………………………What time of day? ? Morning ? Afternoon ? Evening ? Night ? Don’t knowWhat time of year was it? ? Rainy season ? Dry season ? Before Easter ? After Easter ? Don’t know Date (if known) …………………………………………………………………….Weather conditions: ? Dry ? Raining ? Don’t knowWhere did they see the snake? ? Garden ? Plantation ? In the Bush ? In the Rainforest ? Stream ? Don’t know ? Other………………………………………….Were there many plants covering where they were walking?? Yes- thick coverage ? Yes- medium coverage ? No- little or no coverage? Don’t knowWere they using a footpath/ road? ? Yes ? No ? Don’t knowHow close did they get to the snake? ? 0-1 metres ? 1-2m ? 2-5m ? 5+ m ? Don’t knowDid they get a good view of the snake?? No- saw it very briefly for only a couple of seconds ? No- it was partly hidden ? Yes- clearly viewed the snake for several seconds ? Yes- watched the snake for a long time ? Don’t know What was the snake doing? ? Moving on ground ? Resting on ground ? Moving in tree ? Resting in tree ? Don’t know ? Other …………….................What precautions against snake-bite did they take?? Wearing protective clothing/ footwear ? Treading carefully ? None ? Use a torch (if at night) ? Take a dog ? Don’t know ? Other ………………………………………………………….If wearing protective clothing, what was this?? Boots ? Long trousers ? Knee length “snake-gaiters” ? Gloves? Thigh length “snake-chaps” ? Don’t know ? Other ………………………What activity were they doing?? Gardening ? Recreation ? Farming ? Hunting/ collecting plants ? Working in Forest ? Don’t know ? Travelling? Other…………………………………………………………..Did they try to kill this Fer-de-Lance? ? Yes ? No ? Don’t knowDid they succeed in killing it? ? Yes ? No ? Don’t knowIf yes, what did they kill it with? ? Metal pole ? Cutlass ? Rock ? Gun ? Spear ? Stick ? Don’t know ? Other ……………………..Did they keep any of the body parts? ? Yes ? No ? Don’t knowIf yes- what parts were they? ......................................................................................What were they used for? ..........................................................................................What are their reasons for wanting to kill the snake? ……………………………………………………………………………………………………………………………………………………………………………………Has a Fer-de-lance ever bitten/ tried to bite them? ? Yes- successful bite ? Yes- unsuccessful bite ? Don’t know ? NoIf yes- where was the geographical location? .............................................................. ? Don’t knowGPS……………………………………………………………………………………..What time of day? ? Morning ? Afternoon ? Evening ? Night ? Don’t knowWhat time of year was it? ? Rainy season ? Dry season ? Before Easter ? After Easter ? Don’t know Date (if known) …………………………………………………………………….Weather conditions: ? Dry ? Raining ? Don’t knowWhere was the snake? ? Garden ? Plantation ? In the Bush ? In the Rainforest ? Stream ? Don’t know ? Other………………………………………….How large was the snake? ? Small (Juvenile) ? Large (Adult) ? King-Size (very large adult) ? Don’t know Were there many plants covering where they were walking?? Yes- thick coverage ? Yes- medium coverage ? No- little or no coverage? Don’t knowWere they using a footpath/ road? ? Yes ? No ? Don’t knowWhat precautions against snake-bite did they take?? Wearing protective clothing/ footwear ? Treading carefully ? None ? Use a torch (if at night) ? Take a dog ? Don’t know ? Other ………………………………………………………….If wearing protective clothing, what was this?? Boots ? Long trousers ? Knee length “snake-gaiters” ? Gloves? Thigh length “snake-chaps” ? Don’t know ? Other ………………………What activity were they doing?? Gardening ? Recreation ? Farming ? Hunting/ collecting plants ? Working in Forest ? Don’t know ? Travelling? Other…………………………………………………………..Did they see the snake before it bit them? ? Yes ? No ? Don’t knowIf No- Did they tread on the snake? ? Yes ? No ? Don’t knowDid they try to kill the snake, try to move it away or try to get close to it? ? Yes ? No ? Don’t knowIf they did try to kill the snake, were they successful? ? Yes ? No ? Don’t knowDid the snake attack them first before they tried to kill it? ? Yes ? No ? Don’t knowWhat are their reasons for wanting to kill the snake? ……………………………………………………………………………………………………………………………………………………………………………………What did they kill it with? ? Metal pole ? Cutlass ? Rock ? Gun ? Spear ? Stick ? Don’t know ? Other ……………………..Did they keep any of the body parts? ? Yes ? No ? Don’t knowIf yes- what parts were they? ......................................................................................What were they used for? ..........................................................................................Where was the bodily location of the snake bite? ......................................................Did they seek medical treatment after being bitten? ? Yes ? No ? Don’t knowIf No- what were the reasons for not seeking medical treatment? ....................................................................................................................................................................................................................................................................................Did they try to treat the bite themselves or use any traditional medicines/ treatments?? Yes ? No ? Don’t knowHow did they treat their bite/ what traditional medicine did they use?………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………Where did they hear about this treatment? ....................................................................................................................................................................................................................................................................................What were their injuries? ………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………How long did they take to recover? ............................................................................Did they know anybody else who had been bitten? ? Yes ? NoIf yes- What are their contact details? ……………………………………………………………………………………………………………………………………………………………………………………? Don’t knowWhat is the respondents’ relationship with this person?? Family (this respondent is the victims..……………) ? Friend ? Neighbour? Colleague ? Don’t know the victim personally ? Other………………..Have any of their animals/ livestock/ pets been bitten by Fer-de-Lance?? Yes ? No ? Don’t knowWhat type of animal was this?? Dog ? Cat ? Pig ? Goat ? Sheep ? Horse ? Poultry? Don’t know ? Other …………………………………….Where was the geographical location? ........................................................................Did the animal survive? ? Yes ? No ? Don’t knowWhat are their feelings of Fer-de-Lance?? Strongly dislike ? Dislike ? Neutral (neither like or dislike) ?Like ? Strongly like ? Don’t knowWhat are the reasons for this? ………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………………Do they think there is anything good about Fer-de-Lance? ? Yes ? No ? Don’t knowIf yes- What? ……………………………………………………………………………………………………………………………………………………………………………………Do they know any traditional snakebite treatments? ? Yes ? No ? Don’t knowIf yes- what are they? ……………………………………………………………………………………………………………………………………………………………………………………How would they behave if they encountered a Fer-de-Lance?? Try to kill it ? Try to move it or chase it away ? Panic and run away? Walk away calmly ? Don’t know ? Other ……………………………..What would they do if they got bitten?? Seek medical help ? Use traditional medicine ? Do nothing ? Don’t know? Other ……………………………………………If the Fer-de-Lance became a protected species, would they feel differently about the snake?? Yes ? No ? Don’t knowDo you think numbers of Fer-de-Lance have changed? ? Yes- increased ? Yes- decreased ? No ? Don’t knowIs it possible to tell the difference between Fer-de-Lance and a Boa Constrictor?? Yes ? No ? Don’t knowHow would you tell the difference?? Shape of the head ? Length of snake ? Thickness of snake ? Colour ? Behaviour ? Pattern on skin ? Shininess ? Other ………………………….Do feral pigs and Fer-de-Lance live in the same areas? ? Yes ? No ? Don’t knowDo you think mongoose help control Fer-de-Lance? ? Yes ? No ? Don’t knowHave you ever seen a mongoose eat a Fer-de-Lance? ? Yes ? No ? Don’t knowAdditional notes …………………………………………………………………………………………Appendix 2: Hospital records formToday’s date ………..……. Number of records searched ………………..Docket No. (Case notes No.) ……………………………………………………….Gender: ? Male ? Female Patients age ………………………… ? Child ? Adult up to 40 ? Adult between 40-60 ? Adult over 60 Patients Occupation: …………………………………………………………………Date of bite ……………………………………………… Year …………………….Date of hospital admission ……………………………………………………………Geographical location of snake bite ………………………………………………….Bodily location of snake bite: ? Foot ? Lower leg ? Upper leg ? Hand ? Lower arm ? Upper arm ? Torso ? Head/ neck ? Other ……………………………Outcome? Loss of limb ? Loss of use of limb ? Full recovery- no long-term damage ? Scarring ? Death ? Other ……………………………………………….… meow...d r ... Duration of hospital treatment ? Less than 1 week ? 1-2 weeks ? 2-3 weeks ? 3-4 weeks ? 4-6 weeks ? 6+ weeks******************************************************************************** ................
................
In order to avoid copyright disputes, this page is only a partial summary.
To fulfill the demand for quickly locating and searching documents.
It is intelligent file search solution for home and business.
Related searches
- school home page elementary
- xfinity home page install
- xfinity home page internet explorer
- centre for globalization research ca
- global research centre for globalization
- fwisd home page for students
- xfinity home page or homepage
- xfinity home page install for windows 10
- centre for research on globalization
- conservation science and technology
- global research centre for research on globalization
- centre for financial inclusion