Review: Welfare Outcomes Of Leg-Hold Trap Use In Victoria
Prepared By Nocturnal Wildlife Research Pty Ltd September 2008
View the PDF version of this document: Welfare outcomes of leg hold trap use part 1
Nocturnal Wildlife Research Pty Ltd (NWR) is a business committed to innovative research, development and education in the field of vertebrate pest control and wildlife management. Nocturnal Wildlife Research Pty Ltd takes pride in the accuracy of its material but does not guarantee that this document is without flaw of any kind and is wholly appropriate for your purposes and therefore disclaims all liability for error, loss or other consequence that may arise from you relying on any information in this publication.
Peer Review Process
The information contained in this review has peer reviewed by experts external to NWR and scientific advice and comment has been sought where appropriate and with reference to the needs of the client. This does not guarantee that this document is without flaw of any kind.
Potential Conflict Of Interest Disclosure
Mention of trade names is for identification purposes only and does not constitute endorsement or disendorsement by NWR. Readers should note that a ‘potential conflict of interest’ describes a circumstance or association, not a behavior, and does not imply that one exists. In its normal course of business activities NWR adopts a policy of ’transparency’ in such matters that might be perceived as interfering with objectivity and impartiality in dealing with clients who request scientific advice. This document has been prepared with due regard to objective assessment and interpretation. Readers should be aware that NWR has an open commercial interest in developing the Tranquilliser Trap Device (TTD) and Lethal Trap Device referred to in this document.
Citation System In This Document
The senior author is cited in the text with the date of publication followed by ‘et al.’ to denote one, or more than one, co-author. This system was used to save page space and to facilitate easier incorporation of citations into tables.
|(i)||GLOSSARY OF ABBREVIATIONS AND TERMS|
|1.0||AIMS, OBJECTIVES AND METDODS|
|2.1||Use of leg-hold devices in Victoria|
|2.2||Traps and snares used world-wide for canid control|
|2.2.4||Cage (or box) traps|
|2.3||World-wide regulation of leg-hold traps|
|2.4||Limitations of trapping as a control technique|
|3.0||DEFINING WELFARE OBJECTIVES|
|3.1||What is ‘good welfare’ and can we recognise it?|
|3.2||Humane vertebrate pest control|
|3.3||What is a humane trap?|
|4.0||IDENTIFICATION OF TARGET AND NON-TARGET SPECIES|
|4.1||Defining target and non-target species|
|4.2||Common non-target species in south-eastern Australia|
|4.3||Discussion and conclusions|
|5.0||IDENTIFYING INDICATORS OF TRAPPING STRESS|
|5.1||Stress and stressors|
|5.4||Visible pathological indicators|
|5.5||Survival, growth and development|
|5.6||Discussion and conclusions|
|6.0||STRESSORS AND PATDOLOGY ASSOCIATED WITD LEG-HOLD|
|6.1.2||Primary trauma and acute pain|
|6.1.3||Restraint and handling|
|6.1.4||Behavioural, social and spatial dislocation|
|6.1.5||Loss of cover|
|6.1.8||Food and water|
|6.2.1||Secondary trauma and pain|
|6.2.2||Anxiety and fear|
|6.2.3||Capture myopathy and exhaustion|
|6.2.4||Hyperthermia and hypothermia|
|6.2.5||Impact upon dependent young and reproduction|
|6.2.6||Dehydration and starvation|
|6.3||Discussion and conclusions|
|7.0||COMPARISON OF DEVICES|
|7.1.1||Steel-jawed leg-hold traps|
|7.1.2||Modified steel-jawed leg-hold traps|
|7.1.3||Padded leg-hold traps|
|7.1.4||Laminated leg-hold traps|
|7.2||Comparative capture rate|
|7.3||Comparative capture efficacy|
|7.5||Discussion and conclusions|
|8.0||METDODS TO IMPROVE WELFARE OUTCOMES|
|8.1||Assessing trap performance|
|8.2||Trap inspection times|
|8.4||Deactivation of traps|
|8.6||Trap size and weight|
|8.8||Tranquilliser trap device (TTD)|
|8.9||Lethal trap device (LTD)|
|8.10||Trap signalling devices|
|8.11||Lures, odours and attractants|
|8.12||Euthanasia or release?|
|8.13||Trap sets and target-specificity|
|8.15||Jaw off-set distance|
|9.0||GENERAL CONCLUSIONS AND RECOMMENDATIONS|
|9.2||Definition and regulations of leg-hold devices|
|9.3||Development of trap specifications|
|9.4||Improving welfare outcomes|
|9.6||Assessing comparative welfare outcomes|
|9.7||Reporting research and assessment|
|1.0||Haematological and biochemical responses of red foxes (Vulpes vulpes) to|
|different recovery methods (C.A. Marks, in review)|
|2.0||List of steel-jawed, padded and snare leg-hold devices|
|3.0||Trapping practices used for canid research in Australia|
Glossary Of Abbreviations And Terms
Animal Ethics Committee
|Anxiety||Prolonged apprehension or worry that may affect mood, behavior and physiological activity|
|Autotomy||Removal (usually by biting or chewing) by an animal of its own appendage as a means to escape and self-directed trauma in response to nerve injury|
|Canid||Member of the family ‘Canidae’ (eg. dogs and foxes)|
|CK-MM||sub-fraction Isoenzyme sub-fraction of CK expressed by skeletal muscles|
|CNS||Central Nervous System|
|dB||Decibel – a unit of relative sound loudness|
|Exotic pest||A species translocated from a foreign ecosystem now existing in a free-living (wild) state that is considered to negatively impact upon a particular resource or value|
|Feral pest||A once domesticated species now existing in a free-living (wild) state that is considered to negatively impact upon a particular resource or value|
|Foot-hold trap||A trap employing two jaws hinged and held open by a trigger mechanism that when stepped on closes by spring action around the foot or leg, preventing the animal from escaping. Some overseas trapping standards and commercial literature define a foot-hold trap as one where the ‘jaw spread’ is less than six inches. In this review the terms leg-hold and foothold are used interchangeably as specified by the authors cited and with reference to Victorian legislation that makes no distinction between leg-hold and foot-hold traps|
|Fourth (4th) generation||A padded (‘rubber jawed’) trap manufactured by Woodstream Corporation in Pennsylvania|
|Victor Soft-Catch trap||(USA) that has been progressively modified since its initial versions were reported in published studies in the early 1980s. Published accounts since the early 1990s assess ‘4th generation’ versions that are reported to have greater efficacy and different attributes compared to previous trap generations|
|Ischemia||A decrease in the blood flow to a tissue or organ|
|ISO||International Organisation for Standardization|
|Jaw spread||Distance between opposing jaws of leg-hold (or foot-hold) traps when open and set|
|Laminated steel jaw traps||Flat jaw (no teeth or serrations) style leg-hold trap, with metal added to the jaws to increase their surface area|
|LTD||Lethal Trap Device|
|Macropods||Member of the family Macropodidae (eg. kangaroos and wallabies)|
|MCV||Mean corpuscular volume|
|Leg-hold trap||A trap employing two jaws held open by a trigger mechanism that when stepped on closes by spring action around the foot or leg, preventing the animal from escaping. Some overseas trapping standards and commercial literature define a leg-hold trap as one where the ‘jaw spread’ is greater than six inches. In this review the terms ‘leg-hold’ and ‘foot-hold’ are used interchangeably as specified by the authors cited and with reference to Victorian legislation that makes no distinction between leg-hold and foot-hold traps|
|mg kg-1||Milligrams per kilogram|
|N||Newtons – a unit of force|
|N:L ratio||Neutrophil to lymphocyte ratio|
|Non-target species||Animals that are not the target of control and are ‘by-catch’ or affected unintentionally|
|NT:T ratio||Non-target to target species ratio|
|Oedema||The presence of abnormally large amounts of fluid in the intercellular tissue spaces|
|Pain||An unpleasant sensory and emotional experience associated with actual or potential tissue|
|damage or described in terms of such damage|
|PCV||Packed cell volume|
|RBC||Red blood cell|
|RCC||Red (blood) cell count|
|Self-mutilation||Used to describe self-inflicted bite wounding, usually of trapped limb|
|Serrated steel jaw trap||Tooth-style leg-hold trap made from steel and spring operated by pressure applied to a plate|
|or treadle in the centre of the device|
|Soft jaw or rubber jaw||Flat jaw (no teeth) style leg-hold traps which have rubberised padding added to the jaws and|
|traps||is spring operated by pressure applied to a plate or treadle in centre of the device|
|Target species||Animals that are the target of control and captured or affected intentionally|
|Treadle-snare||Tennis racquet-shaped device, spring arm operated by pressure applied to a plate or treadle in centre of the device which pulls tight on a cable that may be plastic/rubber coated that snares the animal|
|TS||Trap selectivity – one measure of the ‘target-specificity’ of a trap|
|TTD||Tranquilliser Trap Device|
|WBC||White blood cell|
|WCC||White cell count|
|WTO||World Trade Organisation|
- In a significant portion of the current distribution of sheep and cattle in Australia, the dingo (Canis lupus dingo) and its hybrids (generically known as wild dogs) are implicated as a predator of livestock (predominantly sheep). Surplus killing behaviour of dingoes may result in a large number of livestock deaths and wounding.
- In Australia, trapping and leg-hold snaring is mainly used for wild dog control in locations where shooting or poison baits are deemed to be inappropriate given proximity to settlements, where baiting has little impact or where legal restrictions are imposed on where baits may be laid.
- Leg-hold traps (and snares) have received much attention from animal welfare and anti-trapping lobby groups worldwide over poor welfare outcomes. The leg-hold trap was banned in the UK in 1958 and is now banned in at least 80 countries. Restriction on the use of many leg-hold traps will commence in New Zealand by 2011. In Victoria, the use of large steel-jawed (eg. Lane’s) leg-hold traps for wild dog control is still authorised in proclaimed exclusion zones.
- The purpose of a restraining trap (or snare) is to reliably capture and hold the animal unharmed with the minimum of stress until the trap is checked and the animal can be euthanased or released. Overall welfare of the target and non-target species from the moment of capture until intervention due to euthanasia, death through other causes or after release from the trap is relevant to the overall and relative humaneness of traps.
- Traps with the best relative humaneness will minimise suffering and permit an acceptable balance of the harms associated with trapping against the benefit of effective trapping of wild dogs.
- Animals that are captured unintentionally by traps are commonly referred to as ‘nontarget’ species. A trap is considered to be more selective if it captures a higher proportion of “target” species rather than wildlife, domestic or exotic animals that are incidental to the objectives of the control programme. Other factors that affect trap selectivity include the location and manner in which it is set and the attractants used.
- A reduction in the capture of non-target species implies a corresponding reduction in negative welfare impacts that have no beneficial outcome. If few traps are occupied by non-target species, there is a greater potential for the capture of target species.
- Common wombats, swamp and red-necked wallabies, brushtail possums and eastern grey kangaroos are very common non-target species taken by leg-hold traps in southeastern Australia, along with exotic non-target species such as the red fox, feral cat and European rabbit. Superb lyrebirds, goannas, echidnas, emus and corvids are also frequent non-target captures.
- The behaviour of some non-target species may make them susceptible to capture. The common wombat’s propensity to mark areas of disturbance may promote their capture at trap sites prepared by digging, clearing or movement of logs or trap setting at the base of trees.
- Trapping releases predictable physiological responses as a reaction to a range of stressors encountered during capture. Attempts to measure the welfare impact of trapping can be made by measuring the magnitude of the biological response, pre-pathological state and consequent pathology.
- Potentially there may be a wide range of stressors associated with trapping, many of which are not directly related to the trap mechanism. Startle, primary acute trauma and pain, restraint, handling, noise, light, loss of cover, social and spatial dislocation, food, odour, water and thermal stressors may act in various combinations to influence the degree to which animals resist traps and the overall stress and welfare outcomes of trapping.
- Secondary physical trauma (eg. ischemia, predation, insect attack etc), chronic pain, anxiety and fear, self-mutilation, capture myopathy, exhaustion, impacts on young (loss of dependent young, ejection of pouch young and abortion etc), starvation, dehydration, hypothermia, hyperthermia and death are pathological endpoints of stress and the consequence of exposure to intense stressors or a combination of stressors. Good welfare outcomes of trapping should seek to prevent or mitigate such consequences.
- The assessment of injuries using trauma scales to determine welfare is limited in its ability to estimate the impact of many stressors and pathological outcomes of trapping. A key deficiency associated with the use of trauma scoring in trap studies is that the amount of time that an animal spends in captivity is rarely known with any accuracy.
- Data logging systems that reveal the capture time, duration and relative activity of animals are likely, in conjunction with physiological indicators such as CK, AST, ALP, ALT and N:L ratios as well as whole body necropsies, to enable the most useful, practical and unequivocal insights into the relative welfare impacts of traps.
- Many of the haematological and biochemical indicators are standardised, cost-effective and widely available laboratory tests that, if properly applied, could provide sufficient information to monitor relative welfare states and promote adaptive management of trapping practices towards better welfare outcomes.
Comparison Of Devices
- Padding of leg-hold trap jaws has been attempted with cloth, plastic or rubber tubing in an ad hoc manner in a number of overseas and Australian studies. This results in less injury than that produced by unmodified devices, but does not offer superior outcomes compared to those associated with commercially available padded traps.
- International literature suggests that in most cases, leg-hold snares are less effective than leg-hold traps for canid control. Some data suggests that treadle-snares cause greater stress to red foxes than other capture devices and the continued use of the treadle-snare should be reviewed with reference to these new data.
- Laminated leg-hold traps have been found in some studies to reduce the incidence of trap related injury, when compared to similar non-laminated devices. Currently there is no clear scientific consensus that laminated traps have the potential to deliver better welfare outcomes compared to commercially available padded leg-hold traps. Lamination of existing leg-hold traps is unlikely to produce significantly improved welfare benefits compared to padded devices.
- Devices that conform to the ‘fourth generation’ of the Victor Soft-Catch #3 trap probably represent current best practice in canid trapping that can be determined from published information. There appears to be potential for optimal welfare outcomes using commercially available padded leg-hold traps that use short restraining cables, standard pan tension systems, are suited to the attachment of TTDs or LTDs, are more familiar to trappers and are well supported by published efficacy data for the capture of canids.
Promoting Better Welfare Outcomes
- In order to promote current best practice and reliable welfare outcomes, mechanical trap specification should be established that clearly define minimum performance based attributes. Important trap specifications should include trap size and jaw spread, trap weight, closure speed, impact force, clamping force, jaw offset distances, padding material and pan tension characteristics. Ancillary features used with traps such as the type and number of in-line springs, swivels and anchoring methods should also be specified. A minimum benchmark could be based upon the fourth generation Victor Soft-Catch #3 trap using the manufacturer’s data or physical measurements.
- Evaluating trap performance and routine testing and maintenance of traps will reduce the likelihood of failure in the field and poor welfare outcomes that result. The performance characteristics of traps such as spring tensions and closing speed will greatly influence the position on the limb where animals are restrained and the resulting trauma sustained.
- A positive relationship exists between the periods of time held in captivity and the degree of injury and stress. In most countries in the developed world, trap inspection periods of at least once per day are a minimum standard. Nocturnal animals are likely to experience additional stress if held for prolonged periods during the day. In the absence of novel ways to demonstrably improve the welfare of animals held for periods in excess of one day, trap inspection periods should be at least once per day.
- Various studies have contrasting recommendations concerning the merits of anchored or ‘drag’ fixed trap restraints. It would be appropriate to monitor the welfare outcomes, using appropriate scientific protocols, where both options are used for target and non-target animals and adopt the most beneficial practice.
- In-line spring specifications that have been developed in North America are unlikely to have catered for species such as macropods that are capable of developing very large amounts of momentum over short distances. The specification of in-line springs in trap restraining chains should be adequate to ensure that the large forces of momentum produced by macropods (eg. kangaroos and wallabies) and predators are based upon realistic calculations of force that can be produced given the length of the chain, acceleration and their mass.
- Pan tensioning (adjustment of ‘trigger’ sensitivity) is a proven, practical and inexpensive way to increase target-specificity and improve welfare outcomes. It will be most effective if applied to standard trap types and trap setting procedures and based upon empirical studies that seek to understand the most appropriate trigger forces that allow reliable capture of target species and exclusion of non-targets. Regular and standardised assessment of the performance of pan tensioning devices should be undertaken in the normal maintenance of overall trap performance.
- Trap size and jaw spread affects the incidence of non-target captures and is probably an important way to limit capture of macropods and other non-target species. There is no evidence to suggest that capture rates and trap efficacy are significantly reduced by using leg-hold traps that have a reduced jaw area/size.
- Use of Tranquilliser Trap Devices (TTDs) may have significant advantages for increasing the efficacy and welfare outcomes of traps. A Lethal Trap Device (LTD) formulation that causes the rapid death of trapped dogs and foxes may prevent injury sustained soon after capture and prevent the distress of prolonged confinement and/or after debilitation. Both approaches may also reduce the potential for dogs to escape if they are not adequately restrained by the trap.
- Trap monitoring systems may be desirable if they prompt trap attendance soon after capture. Most nocturnal target and non-target species are probably captured during the night. Trap attendance after some hours or after an entire evening of captivity may not greatly increase welfare outcomes as much of the significant trauma will occur within the first few hours (possibly within the first hour) of capture.
- The potential exists for lure/odour compounds to increase the target specificity of carnivore trapping by repelling native herbivores (eg. macropods and wombats) from trap sets. Deterrence of native herbivores would be a major advance to limit the capture of a significant number of non-target animals.
- Practices that are used to release non-target species should be reviewed and appropriate equipment and training needs considered to ensure firstly that criteria for the choice between euthanasia and release are known and secondly that if release is attempted it can be done safely, humanely and in conjunction with simple treatments that could reduce post-capture stress and pathology. Macropods and birds may be highly susceptible to capture myopathy and in the absence of knowledge concerning the existence of this disease, routine euthanasia may be the most appropriate action.
- Existing euthanasia recommendations for the use of firearms are probably inadequate and impractical under some circumstances for a range of non-target species and should be reviewed.
- There is a large potential to adapt and modify trapping devices and practices to increase their effectiveness and produce improved welfare outcomes appropriate for local conditions. However, much of the published literature indicates ad hoc field experimentation with inadequate experimental controls and/or the use of multiple modifications or erratic variations in adaptations of the devices. This does not provide a good scientific basis for assessment and technique development.
- This review concludes with a series of recommendations to promote the adoption of best practice trapping of canids to improve welfare outcomes and foster a culture of continuous improvement.
1.0 Aims, Objectives And Methods
The client has requested that Nocturnal Wildlife Research (NWR) Pty Ltd provide a comprehensive literature review to identify the nature of welfare impacts produced by leg-hold and foot-hold traps (here after referred to as leg-hold traps) with reference to Victorian species, and outline directions to promote improvement in welfare outcomes.
- Identify (target and non-target) species within Victoria that are susceptible to leg-hold trapping and describe the relative incidence of capture from published records;
- Describe the welfare impacts that can be anticipated from the use of leg-hold traps;
- Review the merits of trap types, actions and strategies that have the best potential to mitigate a range of welfare impacts.
A literature review was conducted using CAB abstracts, Web of Science, Biosis, PubMed and Google Scholar search engines. The search focused upon compiling a bibliography for the:
- History, current use, regulation and development of leg-hold traps;
- Welfare impacts of leg-hold traps including trapping stressors, associated pathology and techniques to measure behavioural, pathological, biochemical and haematological indicators of poor welfare and stress;
- Assessment of different trap types in Australia and overseas for their comparative humaneness and welfare impacts.
Identifying target species
A review was conducted for studies of trapping and restraint of dingoes (Canis lupus dingo) and red foxes (Vulpes vulpes) in Australia and published information on the impacts of trapping upon non-target species. Literature relating to the trapping and restraint of coyotes (Canis latrans), wolves (Canis lupus), domestic dogs (Canis lupus familiaris), silver foxes (Vulpes vulpes) and Arctic foxes (Alopex lagopus) in overseas studies were used extensively.
Identifying non-target species
Non-target species in Victoria were identified by reviewing the scientific literature for records of target and non-target captures where a range of leg-hold traps and snares have been used for wild dog and fox control1. The occurrence of these species within the leg-hold trap exemption zone (Figure 1) was investigated by using species distribution data from the Victorian Wildlife Atlas (Department of Sustainability and Environment: Victoria). By combining all trapping studies a pool of 1123 wild dog captures were identified, along with associated non-target captures. The relative incidence of non-target species captured relative to each 100 wild dog captures was used to express the likely susceptibility of various species to capture with leg-hold devices expressed as a subjective score of very common (≥ 10 records per 100 wild dog captures), common (1 – 9 records per 100 wild dog captures) or uncommon (≤ 1 record per 100 wild dog captures).
Haematology and biochemistry data for red fox captures
During 1990 – 1994 routine blood sampling of foxes captured in the urban area of Melbourne (Marks et al. 1998; 1999a; 1999b) was undertaken and haematology and blood biochemistry profiles were produced. Treadle-snares, Victor® Soft-Catch™ and cage traps were used to sample foxes, along with the use of netting. A sample of shot foxes was taken at the end of the study. Relevant haematology and biochemistry data are compared with published data for foxes taken by different trap types and for different durations of captivity and sampling techniques. Details of the analysis are contained in Marks (submitted) (Appendix 1).
1 Contemporary data for the trapping of target and non-target animals was sought from the Department of Primary Industries (Victoria) for this review but it was not provided.
2.1 Use of leg-hold devices in Victoria
In a significant portion of the distribution of sheep and cattle in Australia, the dingo (Canis lupus dingo) and its hybrids (generically known as wild dogs) are implicated as a predator of livestock (Fleming et al. 2001). Sheep (Ovis aries) and goats (Capra hircus) are highly vulnerable to predation by wild dogs primarily due to their ineffective anti-predator strategy of fleeing and mobbing (Allen et al. 2004). The control of exotic red foxes (Vulpes vulpes) is also undertaken to protect lambs and to support the conservation of wildlife species (Saunders et al. 1995). As in other countries, the fox is considered a sporting resource (Reynolds et al. 1996) and a vector of some zoonotic diseases such as echinococcus (Jenkins et al. 1992, Saunders et al. 1995). Wild dogs are believed to cost $AU66.3 million in lost agricultural production and control effort (McLeod 2004) and are a much more significant predator of livestock than foxes. Leg-hold traps have been used for the selective removal of individual dogs that attack livestock, and prior to the development of poison baiting, trapping was the primary means of wild dog control in Australia (Harden et al. 1987). Trapping and leg-hold snaring is currently used for wild dog control in locations where shooting or poison baits are deemed to be inappropriate given proximity to settlements, where legal restrictions are imposed on where baits may be laid (Croft et al. 1992) or where baiting has little impact (Fleming et al. 1998). The surplus killing behaviour by individual or a small number of dingoes may result in a large number of sheep being killed in one attack (Thomson 1992, Allen et al. 2001). Consequently, the targeted trapping of dingoes that have commenced predation of livestock may be an important strategy to limit attacks on properties by individual dogs rather than as a means to produce wide-scale reductions in population abundance.
In Victoria, livestock predation by wild dogs is largely restricted to the eastern highlands, where pastures were established in areas surrounded by a large forest boundary, which contains endemic dingo populations (Fleming et al. 2001). Wild dogs are listed as an ‘established pest animal’ under the Catchment and Land Protection Act 1994 (CALP). Wild dog control is mainly carried out by staff of the Department of Primary Industries (DPI), although some control is also undertaken by private land holders. Under Section 30 of the Domestic (Feral and Nuisance) Animals Act 1994, an owner of animals kept for farming purposes (or an authorised officer) is permitted to destroy any dog found at large in the place where those animals are confined. Section 15 of the Prevention of Cruelty to Animals Act 1986 (POCTA) details the offences and the exempted areas for using both large and small leg-hold traps (Figure 1). The Prevention of Cruelty to Animals Regulations 1997 define a large leg-hold trap as one with a jaw spread not less than 12 cm wide, and a small leg-hold trap as one with a jaw spread less than 12 cm. Other than the term ‘spring operated steel jaw leg-hold trap’ contained in Section 15 of POCTA, there is no specific definition in the legislation that takes into account the recent modifications and newer models of leg-hold traps2. Section 6 of the POCTA Act exempts anything done in accordance with the CALP Act, although leg-hold trapping is not specifically mentioned in the CALP Act.
2 The treadle snare was designated to be the device of choice used by Victorian government trappers since 1987 although large serrated steel-jawed traps were still used concurrently until 2004 under authorisation under the CALP Act. Since 2000 large ‘rubber jawed’ (Lane’s type) traps began to be used in Victoria in small numbers until 2006 when they are were adopted increasingly until the phasing out of treadle snare use by December 2007. After the bushfires in 2003, some 790 rubber jawed traps (Jake and modified Bridger #5) have been purchased by DPI in Victoria (B. Roughead, personal communication).
Figure 1. Areas of Victoria where trapping with steel-jawed leg-hold traps is prohibited (no trapping zone) and the exemption zone (exemption zone) where trapping conducted in accordance with the Catchment and Land Protection Act (1994) is authorised.
2.2 Traps and snares used world-wide for canid control
There are four broad categories of restraining traps and snares that have been used for canid control; leg-hold (or foot-hold) traps3, leg-hold snares, cage (or box) traps and neck snares (Powell et al. 2003, Iossa et al. 2007). Killing traps and neck snares are not used in Australia4 and are primarily confined to North American fur harvesting along with deadfall traps, spring traps, lethal snares, drowning traps and pitfall traps (Powell et al. 2003, Iossa et al. 2007). Snares and deadfall devices have a long history of invention by trappers in North America (Petrides 1946). Novel capture techniques including drive nets have been used to capture wolves in forest habitat in Poland (Okarma et al. 1997) and the use of tranquilliser rifles in North America (Gese et al. 1996) but these have not found wide use in Australia and will not be discussed further.
3The foot is the pedal extremity of a vertebrate animal’s leg (including the tarsus, metatarsus and phalanges). The leg refers to the entire limb used for locomotion in vertebrates, suggesting that a leg-hold trap will restrain animals at any point of the limb. There is no universally accepted, evidence-based definition to distinguish “leg-hold” from “foot-hold” trap, although some commercial literature and North American standards define foot-hold traps as having jaw spreads less than six inches across. In this review the terms are used as specified by the authors of the papers referred to, are not retrospectively defined and are made with reference to Victorian legislation that does not distinguish between foot-hold and leg-hold traps.
4 Although some Australian commercial suppliers advertise certain ‘killing traps’, their legal status in various states is unclear and beyond the scope of this review.
2.2.1 Leg-hold traps
Steel-jawed leg-hold traps have been one of the principal devices used to capture fur-bearing animals world-wide (Payne 1980). They have a long history of use in Europe and North America, particularly in the fur industry after the 1850s, corresponding to the development of mass-produced devices and ongoing experimentation leading to the familiar form of steel-jawed leg-hold traps by the late 1800s (Gerstell 1985). In the United States, leg-hold traps are used for the capture of furbearing mammals, for recreational trapping, pest animal control and as a tool to aid subsistence living in wilderness areas. In the United States they are the primary control measure for coyotes involved in livestock damage (Andelt et al. 1999, Conover 2001). In New Zealand, leg-hold traps have been a major control technique for the exotic brushtail possum (Trichosurus vulpecula) (Warburton et al. 2004).
Research into the development of more humane traps to replace conventional leg-hold traps has been ongoing for over a century (Drahos 1952). More recently, research has expanded as steel-jawed traps have received attention from animal welfare and anti-trapping lobby groups world-wide, over negative welfare outcomes caused by their use (Gentile 1987). The traps that have been most commonly used during the 20th century for wild dog control in Australia are toothed, steel-jawed leg-hold traps, as described by Newsome et al. (1983). These traps have a large jaw spread and are sprung by one or two leaf springs. They are commonly called Lane’s traps5 (Lane’s two springs: Stockbrands Pty Ltd, Western Australia). A range of other devices such as the Oneida #14 traps (one spring: Woodstream Corporation, Pennsylvania), Victor #3 and #4 off-set traps (Woodstream Corporation, Pennsylvania), Montgomery #2 and #3 step-in traps (Montgomery Traps Incorporated, Pennsylvania) have also been used in Australia. A conservative estimate reveals 120 commercial variations of steel-jawed leg-hold devices that have toothed or smooth jaws and at least 15 manufacturers that are primarily based in North America. A list of the major steel jaw trap types and manufacturers is listed in Appendix 2.
‘Laminated’ traps have been designed or modified to increase the width of the trap jaws and the surface area of the jaw face to distribute and displace the energy of the spring as it holds the paw of the captured animal (Hubert et al. 1997). This is often achieved by welding an additional steel bar to the jaw face, which also provides a smooth surface area that reduces lacerations as the animal’s paw moves between the jaws. Increasing the spring tension of the jaws when holding the paw is believed to reduce cutting or sawing movements (Houben et al. 1993, Phillips et al. 1996b). Lamination is primarily a modification of existing trap types that have acceptable capture success and are in wide distribution. The main impetus for these modifications has been the need for the fur industry in North America to meet new trap standards. Additional modifications to laminated trap devices include installing heavier springs, adding centre mounted anchor chains, swivels, shock absorbing coil springs and offseting the jaws (Hubert et al. 1997). The list of steel-jawed leg-hold traps in Appendix 2 denotes trap designs that can be obtained in laminated variations. In a survey of trap types used in Australia, a wide range of commercial ‘laminated’ traps (eg. Duke ™, Jake ™, Bridger ™, Victor® etc) were reported to be used (Nocturnal Wildlife Research Pty Ltd, unpublished data).
A range of steel-jawed leg-hold traps have been modified on an ad hoc basis by placing padding on their jaws, although published specifications and performance assessments appear to be absent. Claims that leg-hold traps are ‘padded’ can be based upon a wide range of modifications to the trap with varying materials and benefit for reducing trauma. Lane’s traps were modified by Harden (1985) and the jaws were padded with polythene piping and offset by ensuring a narrow gap remained between the two jaws when the trap was closed. McIlroy et al. (1986) used Oneida No. 14 jump traps modified by filing away the interlocking spikes on the jaws, and binding the jaws with muslin cloth. Thompson (1992) used padded Lane’s leg-hold traps following the methods described by Newsome et al. (1983) (Appendix 3). The Victor #3 Soft-Catch (Woodstream Corporation: Pennsylvania) was the first commercially manufactured padded leg-hold trap to be widely assessed for its ability to limit capture injuries in coyotes (Olsen et al. 1986, Linhart et al. 1988) and the device has appeared in a number of variations. These traps are made with offset jaws (when closed, a gap of 6–8 mm remains between the jaws) and have a rubberised pad on each jaw that is designed to cushion the impact of the closing jaws on the animal’s limb. The padding also provides a surface that prevents the limb from sliding along or out of the jaws. The trigger force that activates the trap can be adjusted by a bolt on the pan swivels. The #3 trap is predominantly used for capturing wild dogs has a jaw spread of 15 cm, and the smaller #1½ trap for capturing feral cats and red foxes has a jaw spread of 13 cm. The fourth generation Victor Soft-Catch #3 trap has replaceable synthetic rubber jaws and a short 15cm long centre mounted swivel chain as a means to prevent limb damage. The # 3½ EZ Grip trap is a heavier device (Livestock Protection Company, Alpine, Texas) that has been used for the capture of wolves and coyotes. This and the former Victor Soft-Catch represents the only widespread commercially available padded traps that have data published concerning scientific field assessments. In a survey of trap types used in Australia, a wide range of commercial ‘rubber-jawed’ traps were reported to be used, including Duke ™and Jake ™ brands (Nocturnal Wildlife Research Pty Ltd, unpublished data).
5The design is based on the traps originally exported from England to Australia by Henry Lane for the control of rabbits. In 1919 he moved production to Newcastle (NSW) to provide for the demand for subsistence trapping. A padded device for control of dogs is still sold under the ‘Lane’s’ name by Stockbrands Pty Ltd, Western Australia.
2.2.2 Leg-hold snares
A range of leg-hold snares have been developed and used in North America for canid control. The most common include the Novak foot-snare (E.R. Steele Products: Ontario, Canada), Fremont foot-snare (Fremont Humane Traps: Beaumont, Atlanta, USA), Panda foot-snare
(E.E. Lee: Green Mountain Inc.), the Belisle snare (Edouard Belisle: Sainte-Veronique, Quebec, Canada) and the WS-T snare (Wildlife Services Specialists: USA) (see Skinner et al. 1990 for diagrams). The Aldridge trap is a popular snare design for the capture of bears and because of its portability, is used in a range of habitats and applications (Johnson et al. 1980), yet a slow trigger mechanism may increase the number of toe captures and the device cannot be buried (Lemieux et al. 2006). The treadle-snare (Glenburn Motors: Yea, Victoria, Australia) is shaped like a small banjo, has two wire springs and a circular pan or treadle. A wire cable snare is placed around the pan and when triggered the snare is thrown up the animal’s limb and tightened by the springs (Meek et al. 1995, Saunders et al. 1995, Fleming et al. 1998) (Appendix 3). The RL04 is a newer variety of snare developed for bear capture and uses a rubber padded snare that is placed in a PVC cylinder that reduces non-target capture, eliminates hind foot and toe captures and produces minimal tissue damage (Reagan et al. 2002, Lemieux et al. 2006). Most snares use unpadded wire or cable to hold the limb, but recent Kevlar based restraining devices have been used in the UK and have proven successful in the capture of European badgers (Meles meles) with little indication of injury (Kirkwood 2005).
2.2.3 Neck snares
Non-lethal neck snares can be free running so that the noose can relax when the animal stops pulling, or they may be spring operated. The United Kingdom (UK) is one of few European countries where neck snares are permitted, primarily for the capture of red foxes and rabbits for population control (Kirkwood 2005). The Collarum restraint (neck snare) (Green Mountain Inc.: Lander, Wyoming) uses a baited tab pull-arm that triggers a pair of coiled springs, a throw arm that propels a cable loop over the head and neck of coyotes and a stop system prevents the animal from being choked (Shivik et al. 2000, Shivik et al. 2002).
The Gregerson, Kelley and DWRC neck snares have a ratchet system that cause the snare to progressively tighten so that the animal is killed by strangulation; these have been widely used to kill coyotes in the United States (Phillips et al. 1996a). Other snare systems have been made from 0.16 cm diameter cable in a range of designs and are commonly used in predator runs beneath wire fences to kill coyotes by strangulation (Phillips 1996). A power snare that used a spring mechanism to tighten the noose and to strangle red foxes was tested as a possible lethal means of harvesting (Proulx et al. 1990). Commercially available power killing snares included the King (Western Creative Services Ltd: Winipeg, Canada), the Mosher (Mosher: Mayorthorpe, Canada) and Olecko (Olecko: Winipeg, Canada) (Proulx et al. 1990).
2.2.4 Cage (or box) traps
Cage (or box) traps have not been widely used to trap canids and are not regarded as efficient capture devices (Powell et al. 2003), and given their bulk, transport is difficult under field conditions (Way et al. 2002). Way et al. (2002) found that cage traps were expensive, not target-specific and they required a long period of pre-baiting (free-feeding) before they were successful in rural locations. Nonetheless, cage traps have been used with some success to trap urban red foxes in the UK (Baker et al. 1998, Baker et al. 2001) and Australia (Robinson et al. 2001) (Appendix 3), kit foxes (Zoellick et al. 1986) and urban coyotes (Shivik et al. 2005). Trapping injuries from cage traps are minor when compared to corresponding studies that used leg-hold/foot-hold traps, yet coyotes had the potential to injure themselves by biting and throwing themselves against the trap (Way et al. 2002). It is possible that the success of cage traps used for coyotes in urban areas is due to habituation and familiarity in negotiating human made obstacles which makes them more vulnerable than coyotes in rural areas which are difficult to capture in cage traps (Shivik et al. 2005). Some novel cage (box) traps have sought to immediately release hydrogen cyanide gas to rapidly kill captive animals, predominantly to assist in the recovery of ectoparasites (Nicholson et al. 1950), yet these are probably impractical and too hazardous for most pest animal control applications.
2.2.5 Kill traps
Kill traps have been assessed for the lethal harvesting of furbearer species in North America such as mink (Mustela vison) (Proulx et al. 1990, Proulx et al. 1991), fishers (Martes pennanti) (Proulx et al. 1993a; 1993b) and lynx (Felis lynx) (Proulx et al. 1995). The Sauvaggeau 2001-8 (Les Pièges du Quèbec: St Hyacinthe) is a trap with two killing bars powered by torsion springs (Proulx et al. 1994b). The Kania trap (E. Kania: Winlaw, British Columbia) is another lethal trap with a striking bar (Proulx et al. 1993). Other kill traps include the C120 Magnum and Conibear 120 that were developed to quickly render furbearers unconscious and promote a quick death. In trials of the C120 Magnum kill trap to harvest martens, a wide range of other species were taken, including weasels (Mustela erminea), mink, red squirrels (Tamiasciurus hudsonicus), flying squirrels (Glaucomys sabrinas), grey jays (Perisoreus canadensis), and whet owl (Aegolius acadicus) (Proulx et al. 1989). Given the lack of target specificity and the risk of such powerful devices to domestic cats and dogs, their testing has not been pursued for canid control (Proulx et al. 1990, Skinner et al. 1990).
2.3 Worldwide regulation of leg-hold traps
The leg-hold trap was banned in the UK in 1958 under the provisions of the Pest Act (1954) and is now banned in some 80 countries (Fox 2004a, in Iossa et al. 2007)6. A range of leg-hold traps including all long-spring and unpadded double-coil spring traps larger than #1, with the exception of those with the Soft-Catch modification (Warburton et al. 2004) will be prohibited in New Zealand by 2011 under legislation passed in 2007.
The Canadian General Standards Board (Anon 1996) and Agreement on International Humane Trapping Standards (Anon 1997) developed trapping standards that followed the establishment of the Federal Provincial Committee for Humane Trapping (Anon 1981). On the initiative of the Canadian Government in 1987, the International Organisation for Standardisation (ISO) (Princen 2004) produced documents (developed by the Technical Committee -ISOTC191) to assess the safety and capture efficacy of traps (ISO 1999a; 1999b) and standards for the efficacy of killing traps (ISO 1999b). No consensus emerged for determining an acceptable level of injury for restraining traps (Harrop 2000, Princen 2004, Iossa et al. 2007), largely because the trapping industry and animal welfare organisations were divergent when defining a humane trap and it was agreed that the ISO standards would produce testing methodology standards only (Harrop 2000). In 1991 regulations arising from the European Parliament (EEC 1771/94) banned the use of leg-hold traps in the European Community and foresaw a ban on 13 species used for fur products from countries that had not initiated bans on the use of leg-hold traps (Princen 2004). In 1993, Canada proposed that the scope of a Humane Animal Traps Project to be ‘standardisation’ and in 1996 the USA introduced voluntary ‘Best Management Practice’ (BMP) for traps (Princen 2004) under the aegis of the International Association of Fish and Wildlife Services (Anon 2003). Both leg-hold traps and snares have become illegal in certain states of the USA (Way et al. 2002) and by 1999 the European Council prohibited the use of leg-hold traps in 15 member countries (Andelt et al. 1999). Due to a dispute arising with the World Trade Organisation, delays in the implementation of bans led to the establishment of another working group, comprising the USA, Canada (and later Russia) to develop standards to facilitate fur trade (Harrop 2000) and resulted in statements about a range of traps in 1996 for the EC, Canada and Russia (Princen 2004). Although legally non-binding, the parties agreed to develop a set of BMP guidelines for trapping, developed by scientific studies in order to reduce pain and discomfort in target furbearers (Andelt et al. 1999).
2.4 Limitations of trapping as a control technique
Populations of wild dogs that have been subjected to recurrent trapping may become increasingly difficult to capture and trapping effectiveness may diminish over time. In northern California, after sustained trapping for many decades, the trapping effort to capture a single coyote was 10 times that required in southern Texas and this was believed to be the consequence of trap shyness (Sacks et al. 1999). It is difficult to measure the degree of trap shyness without an independent means to assess the number of animals that have avoided trap sets.
Large-scale lethal control of dingoes may not always reduce calf losses as livestock loss is not always obviously related to the abundance of dingoes on a property (Allen et al. 1998) and may be unrelated to the density of wild dogs in an area overall (Fleming et al. 2006). In Victoria, the primary method used to reduce wild dog attacks on livestock is trapping within 3-5 km external to private land boundaries with government land and on private properties.
The effectiveness of this approach remains largely untested (Fleming et al. 2006) and departs from the strategy used in most other states where aerial baiting is the predominant control technique. Trapping is not generally considered to be a control technique that can be used cost-effectively and unilaterally to suppress populations or maintain low population abundance of canids (Fleming et al. 1998). It was recommended that buffer control zones for baiting in the semi-arid Pilbara district of Western Australia should be 15-20 km wide, although the results from the study by (McIlroy et al. 1986) and (Harden 1985) indicate that wild dogs living more than 12-20 km inside National Parks in south-eastern Australia are unlikely to move out onto adjacent private land. Newsome et al. (1983) suggest that a zone 3 km wide is probably adequate for SE Australia although trapping is primarily limited to areas within properties in other parts of Australia (Fleming et al. 1998, Fleming et al. 2006).
Allen et al. (1998) suggest that the disruption of stable dingo packs causes a reduction in the size of packs and the number of experienced hunters that kill larger, more difficult prey. By sharing the cost of chasing, attacking and killing prey, dingoes increase their hunting efficiency (Allen et al. 1998, Allen et al. 2001) and group size increases hunting efficiency by sharing the physiological costs of chasing and attacking prey. Dingoes are known to switch between prey species and may alter their social structure in doing so. For instance, in smaller packs or as solitary hunters, dingoes switch from group hunting to hunt smaller mammals (Corbett 1995). Areas subject to lethal control are typically re-invaded with low ranked members of packs with reduced hunting ability that are more likely to target livestock. In the USA, breeding coyotes were most likely to kill sheep, yet trapping efforts appeared to be least effective at targeting them compared to non-breeding animals that caused the least damage (Sacks et al. 1999). Diminished dingo populations may also permit the invasion or expansion of red fox, feral cat (Felis catus) and European rabbit (Oryctolagus cuniculus) populations (Glen et al. 2007), although the magnitude and importance of such impacts are as yet not fully established. However, it may become increasingly necessary to assess the short-term and localised gains of wild dog control in context with wider ecological impacts that may have more significance to agricultural industries.
6 The UK regulations, similar to those in Victoria, appear to make no distinction between ‘leg-hold’ and ‘foothold’ traps.
3.0 Defining Welfare Obectives
3.1 What is ‘good welfare’ and can we recognise it?
If an animal is having difficulty coping with its environment its welfare can be regarded as being poor (Broom et al. 1993). The ‘magnitude’ of an animal’s welfare is generally associated with the incidence, severity and duration of a negative state (Webster 1998) and the capacity of the species to suffer (Littin et al. 2005). Good animal welfare can be described in terms of physical health and positive emotions, such as pleasure and contentment, while poor welfare comes from ill-health, injury, disease and negative emotions such as frustration or fear, which may be described as ‘suffering’ (Dawkins 2006). When assessing the welfare needs of a species it is inappropriate for a label of a pest to automatically imply that poor welfare outcomes are justified (Marks 1999, Morris et al. 2003, Jordan 2005, Littin et al. 2005). The suffering of wildlife is a relevant concern to wildlife managers and the community (Schmidt et al. 1981), although the term animal welfare is frequently confused with a political movement and not as a discipline that attempts to reduce suffering in animals and investigate their welfare states (Schmidt et al. 1981). A range of authors outline the need to reduce suffering inflicted on animals by trapping (Payne 1980, Schmidt et al. 1981) and to ensure that wildlife management practices are not insensitive to animal welfare concerns (Schmidt et al. 1981, Decker et al. 1987, Andelt et al. 1999).
If animals have a conscious experience of negative states (Mendl et al. 2004) they have the capacity to perceive poor welfare states by awareness of feelings, sensations and thoughts (Block 1998). However, the existence of cognitive capabilities that humans identify with is not a reliable indicator of conscious experience in non-human animals (Dawkins 2001b) and it may be easy to overlook suffering that is not relevant to human experience. As humans commonly equate ‘intelligence’ with the capacity to suffer we are generally more concerned about poor welfare in species such as primates and cetaceans (Marino 2002). However, even when other species such as corvids (eg. ravens and crows etc) show high levels of complex cognition which demonstrate: reasoning, flexibility, imagination, prospection and use of tools (Emery et al. 2004) there is usually much less concern for their welfare.
The adaptive benefits of the potential to suffer has a probable evolutionary significance in promoting avoidance of dangerous environments and circumstances that may produce trauma (Dawkins 1998). ‘Human-like’ consciousness is not necessary for the experience of both the sensory and emotional components of pain (Jordan 2005). It is generally accepted that all classes of vertebrates (with the possible exception of fish) perceive pain (Bateson 1991). The relevance of various stressors and the behaviours that they elicit in non-human animals are not directly apparent to humans, nor can we directly perceive poor welfare states in nonhuman animals or have insight into their mental state or perception from direct observation alone (Rushen 1996). Even in human patients it is difficult to interpret the significance of behaviour associated with the perception of pain and other forms of suffering, especially if that patient cannot communicate (Hackman 1996). Pain perception and suffering in nonhuman animals may be influenced by different mental states (Nagal 1974, Harrison 1991, House 1991) that are related to divergent brain function (Bermond 1997). While there may be a range of complex behavioural differences between species in the display of ‘pain behaviours’, it is thought that many species perceive threshold and tolerance limits of pain in a similar way to humans (Cooper et al. 1986), though our ability to easily recognise this and other forms of suffering in non-human animals makes the assessment of welfare states challenging.
3.2 Humane vertebrate pest control
Humane vertebrate pest control requires the selection of feasible control programs and techniques that avoid or minimise pain, suffering and distress to target and non-target animals (RSPCA 2004) and is based upon a simple precept that an animal’s welfare is good in the absence of suffering (Littin et al. 2004). Until comparatively recently, the humaneness of control techniques used for vertebrate pests has received little attention in Australia (Jones 2003, Marks 2003). Increasingly it is accepted that no technique used to kill or manage pest species should cause unnecessary suffering (Scott 1976, Payne 1980, Schmidt et al. 1981, Ross 1986, Fisher et al. 1996, Marks 1999, Jones 2003, Marks 2003, Littin et al. 2004, Littin et al. 2005). Ideally, pest control methods should be effective and easy to use, safe for humans, humane, target-specific, cost effective and environmentally friendly (Marks 1999). There are few examples of pest control methods that achieve this ideal and the selection of pest control agents often require that a compromise be made.
An ethical basis for pest control firstly requires that control is necessary and can be justified, and that the aims can realistically be achieved and measured (Putman 1995, Marks 1999, Jones 2003, Littin et al. 2004, Littin et al. 2005). In Victoria, wild dog trapping is undertaken to manage livestock predation and the welfare implications of stock predation are significant (Allen et al. 2001, Fleming et al. 2001, Allen et al. 2004), yet the most humane control techniques possible should be used to minimise suffering and balance the harms and benefits of such control (Putman 1995, Marks 1999, Marks et al. 2000, Morris et al. 2003, Littin et al. 2005). In other areas of animal use, clear guidelines promote a reduction in animal suffering. Regulation of animal experimentation demands that if the existence or nature of pain or distress experienced by an animal is unknown, or conclusive evidence does not exist to the contrary, an assumption must be made that pain and distress could be perceived (Anon 2007). Moreover, investigators should assume that procedures that could cause pain and distress in humans are likely to cause pain and distress in other animals (Stafleu 2000). In addition, actions should be governed by an assumption of the worst possible outcome, and the cause of the suffering experienced (eg. as one of or a combination of pain, illness or stress) should be given equal weight (Stafleu 2000).
3.3 What is a humane trap?
Very few restraining traps that are used for wildlife species have been tested against agreed standards for animal welfare (Powell 2005) and there remains widespread confusion about what constitutes a ‘humane trap’ and how it should be defined (Harrop 2000). Traps may be more humane than other devices or acceptably humane, yet a humane trap would be one that avoids subjecting an animal to appreciable stress and avoids compromising its welfare in a significant way. Given the significant stress associated with the capture or restraint of wild animals using any known technique, it is unlikely that the development of a truly humane trap will be realistic objective using contemporary technologies.
In North America, humane trap standards are subject to commercial considerations of harvesting fur and the need to conform to restrictions imposed by fur importing countries. Where traps are set for the purpose of wild dog control in Victoria, trapping is conducted to protect the welfare and viability of livestock. The purpose of a restraining trap (or snare) in this instance is to hold the animal unharmed with the minimum of stress until the trap is checked and the animal can be euthanased or released (Iossa et al. 2007). The overall welfare of the target and non-target species from the moment of capture until intervention due to euthanasia or debilitation or death after release from the trap is relevant to deciding the overall relative humaneness of traps. Proulx et al. (1994b) suggested that the definition of a humane live-trap for furbearers should be a trap that is capable with 95% confidence of holding ≥ 70% of animals for 24 hours without serious injury. In North America, benchmarks or thresholds proposed to certify traps as acceptably humane, typically define a proportion of animals (ie. 20-30%) where poor welfare outcomes are acceptable and the welfare of non-target animals is not considered (Harrop 2000, Princen 2004, Harris et al. 2007, Iossa et al. 2007). Accordingly, traps can be deemed acceptable irrespective of a potential to capture and injure a large proportion of non-target species. Australian guidelines for acceptable welfare outcomes and humane treatment of animals do not ascribe thresholds that accept poor welfare outcomes for a proportion of a specified population in experimentation, agriculture, wildlife or companion animal regulations (eg. Anon 2007).
In this review the trap that has the best relative humaneness will be one that minimises suffering and permits a balance of the harms associated with trapping against the benefits of effective trapping of wild dogs.
4.0 Identification Of Target And Non-Target Species
4.1 Defining target and non-target species
Animals that are captured unintentionally by traps are commonly referred to as ‘non-target’ species. A trap is considered to be more selective if it captures a higher proportion of ‘target’ species rather than wildlife species or domestic animals. Trap selectivity (TS) is a measure of the number of non-target animals captured relative to the number of target animals (Newsome et al. 1983) or the number of non-target animals captured relative to a set number of trap nights (Fleming et al. 1998) where a relatively higher value for TS indicates lower selectivity.
Reducing the number of non-target animals captured has two important benefits for a trapping programme. Firstly, if few traps are occupied by non-target species, there is a greater potential for the capture of target species and a reduction in unproductive maintenance of traps. Secondly, if trap selectivity can be increased, a reduction in the capture of non-target species implies a corresponding reduction in negative welfare impacts that have no beneficial outcome. These benefits are complementary and suggest that trap selectivity is a key component in fostering an efficient trapping programme with optimised welfare outcomes.
Incidental capture of exotic pest species (eg. feral cats, European rabbits, and hares) is sometimes reported to contribute to a tally of target captures (Stevens et al. 1987, Murphy et al. 1990, Fleming et al. 1998). Best practice management of vertebrate pests stipulates the importance of defining clear management objectives, options and strategies that focus upon the mitigation of the impact of particular pests upon stated values (Braysher et al. 1998, Fleming et al. 2001). Stating the target animals sought is an important part of defining the aims and objectives for a control programme. Unintentional capture of exotic or feral species not regarded to be primary targets should be identified as ‘exotic’ or ‘feral’ non-target species, although there may be instances where more than one target species is sought (eg. Meek et al. 1995). Liberal definition of target species as the sum of all pest or exotic species will overstate the specificity and effectiveness of a trap and will not assist in the selection of the most appropriate trapping device or technique for specific objective.
4.2 Common non-target species in south-eastern Australia
In a review of trapping records in six locations in eastern Australia, Fleming et al. (1998) listed a range of non-target species including echidnas (Tachyglossus aculeatus), goannas (Varanus spp), wombats (Vombatus ursinus), possums and sheep. Captured birds included ravens (Corvus spp.), magpies (Gymnorhina tibicen) and pied currawongs (Strepera graculina). Newsome et al. (1983) listed many of the above non-target species during the capture of dingoes in north-eastern NSW with the addition of feral pigs (Sus scrofa), red-necked wallabies (Macropus rufogriseus), cattle (Bos taurus), farm dogs (Canis lupus familiaris), emus (Dromaius novaehollandiae), wedge-tailed eagle (Aquila audax), hawks (family Accipitridae), wonga pigeons (Leucosarcia melanoleuca), tawny frogmouth (Podargus strigoides), superb lyrebird (Menura novaehollandiae), spotted quail-thrush (Cinclasoma punctatum), white-winged chough (Corcorax melanorhamphos) and blue tongued lizard (Tiliqua spp.). Corbett (1974) reported that between 1966 – 71, in 4796 trap nights (80% set without lures or baits), Victorian government trappers using steel-jawed (Lane’s) traps recovered 13 dingoes and 261 non-target species. Mammalian species caught included the common ring-tail possum (Pseudocheirus peregrinus), the sugar glider (Petaurus breviceps), the greater glider (Petauroides volans), koala (Phascolarctos cinereus), long-nosed potoroo (Potorous tridactylus), deer (Cervus sp) and a marsupial carnivore (probably a spot-tailed quoll [Dasyurus maculatus]) although specific numbers of each species captured were not reported. In another study, nine dingoes were captured using padded Lane’s traps, along with 11 mammals, 7 birds and one reptile (Harden 1985). Meek et al. (1995) captured a total of 54 animals with Victor Soft-Catch #3 traps and treadle-snares in coastal NSW. Non-target species caught in 'Victor' traps comprised Australian raven (Corvus coronoides), magpie, swamp wallaby (Wallabia bicolor), long-nose bandicoot (Peremeles nasuta) and brushtail possums. Non-target species caught in treadle-snares were Australian ravens, pied currawong and an eastern grey kangaroo (Macropus giganteus). Using treadle-snares in subalpine NSW, Bubella et al. (1998) captured Australian ravens, a feral cat (Felis catus) and common wombats. Sharp et al. (2005a; 2005b) list ravens, pied currawongs, magpies, kangaroos, wallabies, rabbits, hares, echidnas, goannas, wombats, possums, bandicoots, quolls and sheep as potential non-target species.
Non-target capture records have often used local names or generic descriptions for animals that do not permit identification to the species level. ‘Wallabies’ are likely to include both swamp and red-necked wallabies in south-eastern Australia, and possibly other smaller macropods. Similarly, ‘bandicoot’ are likely be either the southern brown bandicoot (Isoodon obesulus) or long-nosed bandicoot (Strahan 1984) in Victorian studies. Brushtail possums could be the common brushtail or mountain brushtail possum (Strahan 1984). There are three species of crows and three ravens in Australia and these are difficult to tell apart, although reports of crow and raven captures are likely to be little Australian ravens, given their wide distribution and abundance (Pizzey et al. 1997).
In trapping data accumulated during wild dog control programs from November 1986 to December 1987, a total of 1189 animals were captured with steel-jawed (Lane’s) traps and treadle-snares in Victoria. Native animals accounted for 34% (n=397) with 7.4% (n=88) of all non-target species being common wombats (Murphy et al. 1990). When target species were defined as wild dogs only, 62% of trapped animals were non-target species. Overall, the diversity of non-target species captured reflected those reported in other trapping studies in south-eastern Australia (Table 1).
Newsome et al. (1983) found that large jawed Lane’s traps had far less target specificity (TS [non-target:target] = 4.79) than smaller Oneida #14 traps (TS = 0.92). The reduction in brushtail possum, wallaby and common wombat captures for the Oneida trap was also a strong indication of different device specificity. Newsome et al. (1983) reported that 'Oneida' traps were unlikely to catch large-footed animals such as wallabies and emus. ‘Oneida’ traps did not catch kangaroos and wombats although some were sprung by these species. Trapping was conducted at the same site and during the same season, although it was possible that more care was taken in setting Oneida traps to avoid non-target species (Fleming et al. 1998). In other studies, estimates of TS range from 4.79 – 0.13, but this measure is biased given the use of various setting techniques conducted in different habitats and seasons and different degrees of trapping effort and correspondingly variable sample sizes of target and non-target species (Table 1).
The proportion of non-target:target species recovered in all studies indicates a ratio of (rounded to the nearest whole number) 96:100 for the red fox, 58:100 for common wombats,
49:100 for wallabies (swamp and red-necked combined), 26:100 for feral cats, 19:100 for brushtail and mountain possums combined, and 10:100 for eastern-grey kangaroos. Species that were represented < 10:100 wild dog captures, but ≥ 1:100 included the European rabbit (9:100), superb lyrebird (3:100), raven (2:100), goanna (2:100), emu (2:100) and echidna (1:100). A range of other species was represented in < 1:100 wild dog captures (Table 1).
Table 1. Capture records for exotic mammals and non-target mammals, birds and reptiles from studies (1-7) conducted in south-eastern Australia using Lane’s (L), Oneida #14 (O), treadle-snares (T), Victor Soft-Catch #3 (V) traps or a combination of trap types (C), where wild dogs (D) or foxes (F) were the target species. The non-target species (NT) and target species (T) captured are expressed as a ratio: (NT:T) is the number of non-target mammals, birds and reptiles (non-target exotics included) captured for every 100 target species or their reported occurrence (Y). (1 = Newsome et al. 1983, 2 = Stevens and Brown 1987, 3 = Bubela et al. 1998, 4 = Meek et al. 1995, 5 = Murphy et al. 1990, 6 = Fleming et al. 1998, 7 = Corbett 1974).
|Target species sought||D||D||D||D||F||D/F||D/F||D||D||D|
|EXOTIC AND FERAL
|Wild dog||Canis lupus dingo||95||51||17||22||11||7||920||-|
|Red fox||Vulpes vulpes||118||25||23||17||71||28||7||791||Y||Y||96.2|
|Feral cat||Felis catus||36||4||4||4||1||240||Y||Y||25.7|
|Feral pig||Sus scrofa||6||1||0.62|
|European rabbit||Oryctolagus cuniculus||21||1||77||Y||8.82|
|European hare||Lepus europaeus||1||0.1|
|Bandicoot||P. nasuta or I. obesulus||3||-||1||Y||Y||0.36|
|Brushtail or mountain possum||Trichosurus sp||49||1||2||9||1||151||Y||Y||18.97|
|Common wombat||Vombatus ursinus||69||4||2||3||571||Y||57.79|
|Wallaby||Wallabia bicolor or Macropus rufogriseus||92||2||10||13||1||434||Y||49.15|
|Eastern grey kangaroo||Macropus giganteus||8||1||100||Y||9.71|
|Farm dog||Canis lupus familiaris||1||0.09|
|Whistling kite||Haliastur sphenurus||0||1||0.09|
|Wedge-tailed eagle||Aquila audax||1||1||1||Y||0.27|
|Wonga pigeon||Leucosarcia melanoleuca||7||3||Y||0.89|
|Tawny frogmouth||Podargus strigoides||1||0.09|
|Superb lyrebird||Menura novaehollandiae||16||1||21||Y||3.38|
|Spotted quail-thrush||Cinclostoma punctatum||1||0.09|
|White-winged chough||Corcorax melanorhamphus||4||0.36|
|Australian magpie||Gymnorhina tibicen||1||5||Y||0.53|
|Pied currawongs||Strepera graculina||1||1||1||Y||0.27|
|Blue-tongue lizard||Tiliqua sp||1||0.09|
Table 2. Major non-target (Status as ‘E‘ = exotic or ‘N’ = native) and record of presence in either the eastern (E) or western (W) trap exemption zone (Zone) with body weight (kg) and subjective frequency of occurrence (*** = very common, ** = common, * = uncommon [see methods for explanation]) in combined trapping records (NTf), comparative activity substrates (AS) (T = terrestrial, SC = scansorial, A = arboreal), activity rhythms (AR) (N = nocturnal, C = crepuscular, D = diurnal) and feeding category (FC) (C = carnivorous, I = insectivore, O = omnivore, BH = browsing herbivore, GH = grazing herbivore) and if a meat diet has been confirmed (Y/N) (Strahan 1984, Lee et al. 1985).
|Common name||Species||Status||Zone||Weight kg||Activity||Diet||Authority|
|Bobuck||Trichosurus cainus||N||E||1.5-3.7||***||N||A||N||BH||Y||Menkhorst et al. 2004|
|Brushtail possum||Trichosurus vulpecula||N||E/W||1.5-4.0||***||N||A/T||N||BH||Y?||How 1988, Marks 2001a|
|Common wombat||Vombatus ursinus||N||E/W||20-35||***||M||T||N||GH||N||Menkhorst et al. 2004|
|Eastern grey kangaroo||Macropus giganteus||N||E/W||< 66||***||N||T||N/CR||GH||N||Menkhorst et al. 2004|
|Echidna||Tachyglossus aculeatus||N||E/W||2-7||*||N||T||D||I||N||Menkhorst et al. 2004|
|Emu||Dromaius novaehollandiae||N||E/W||30-45||*||N||T||D||O||Y||Schodde et al. 1990|
|European rabbit||Oryctolagus cuniculus||E/W||1-2.4||***||E||T||N/CR||GH||N||Menkhorst et al. 2004|
|Feral cat||Felis cattus||E/W||2.5-6.5||***||E||T||N||C||Y||Menkhorst et al. 2004|
|Goanna||Varanus varius||N||E/W||< 20||*||N||T/SC||D||C/O||Y||Cogger 2000|
|Little Australian raven||Corvus coronoides||N||E/W||0.5-0.82||*||N||T||D||O||Y||Schodde et al. 1990|
|Long-nosed bandicoot||Perameles nasuta||N||E||0.85-1.1||*||N||T||N||IO||Y||Lyne 1971, McIlroy 1981, Fairbridge et al. 2001|
|Red fox1||Vulpes vulpes||E||E/W||3.5-8.0||***||E||T||N||C/O||Y||Menkhorst et al. 2004|
|Southern brown bandicoot||Isoodon obesulus||N||E/W?||0.4-1.0||*||N||T||N||IO||Y||Fairbridge 2000, Fairbridge et al.
2001, Menkhorst et al. 2004
|Spot-tailed quoll||Dasyurus maculatus||N||E||< 7.0||*||N||T/ SC||N||C/, IO||Y||McIlroy 1981, Belcher 1998, Menkhorst et al. 2004|
|Superb lyrebird||Menura novaehollandiae||N||E||< 1.5||*||N||T||D||I||N||Schodde et al. 1990|
|N||E./W||< 27||***||N||T||N||BH||Y||Edwards et al. 1975, Menkhorst et al. 2004|
1Identification of bobuck and brushtail possums are likely to be easily confused
2Considered exotic non-target species in some wild dog control programmes
Figure 2. (cont.) Distribution of major non-target species within the east and west trapping exemption zones in Victoria: (a) brushtail possum, (b) common wombat, (c) eastern grey kangaroo, (d) echidna, (e) emu, (f) feral cat, (g) goanna, (h) long-nosed bandicoot, (i) bobuck, (j) little Australian raven, (k) red fox, (l) swamp wallaby, (m) red-necked wallaby, (n) southern brown bandicoot, (o) spot-tailed quoll, (p) superb lyrebird and (q) European rabbit. (n) southern brown bandicoot (m) red necked wallaby (p) superb lyrebird (o) spot-tailed quoll (q) European rabbit
Figure 2. (cont.) Distribution of major non-target species within the east and west trapping exemption zones in Victoria: (a) brushtail possum, (b) common wombat, (c) eastern grey kangaroo, (d) echidna, (e) emu, (f) feral cat, (g) goanna, (h) long-nosed bandicoot, (i) bobuck, (j) little Australian raven, (k) red fox, (l) swamp wallaby, (m) red-necked wallaby, (n) southern brown bandicoot, (o) spot-tailed quoll, (p) superb lyrebird and (q) European rabbit.
4.3 Discussion and conclusions
The wide range of non-target species reported for studies using leg-hold traps and snares in south-eastern Australia supports previous conclusions that trapping with leg-hold devices is not highly target-specific (Sharp et al. 2005a; 2005b). A wide range of native species can be considered as non-target species, with common wombats, wallabies (considered as both swamp and red-necked wallabies), brushtail (and bobuck) possums and eastern grey kangaroos appearing as very common non-target species in south-eastern Australia. Common exotic non-target species include the red fox, feral cat and European rabbit.
The capture of non-target species is highly dependent upon the geographical distribution of animals and their population abundance in particular environments (Shivik et al. 2002) and is subject to seasonal and long-term fluctuations. The habitats in which traps are used and the foraging behaviour of animals that bring them into contact with traps influences non-target captures. The manner in which the trap is set, its location (Powell et al. 2003), selectivity of the device used (eg. pan tension settings: see Turkowski et al. ), trap size: see Newsome (1983) and the proportion of animals that are restrained by the trap without escape (Shivik et al. 2002) will determine the measured TS of the device.
Species with reduced distribution or low abundance could theoretically be highly susceptible to some traps, yet may not be well represented in capture records. A table of major non-target species was prepared with the emphasis upon species that were represented in more than
1:100 wild dog captures (Table 2). Although uncommonly represented in non-target capture records, bandicoots and spot-tailed quolls were included as potential non-targets as they are restricted or patchy in distribution and/or exist in low to moderate density in some locations (Figure 2). This could suggest the potential to be a more frequent non-target species in specific locations.
Corvids (eg. crows and ravens) are cosmopolitan and appear to be commonly represented in many trapping studies worldwide. American crows (Corvus brachyrhynchus), common ravens (Corvus corax), grey jays (Perisoreus canadensis) and blue jays (Cyanocitta cristate) were frequently captured in a range of leg-hold traps in Canada, while hawks, eagles and owls were captured less often and ducks (Anatidae) were captured rarely (Stocek et al. 1985). Notably, deer appear to be common non-target species in the United States (Pruss et al. 2002), yet although extensive exotic populations exists in Victoria (Strahan 1984), there was no enumeration of deer captures, other than an unspecified report by Corbett (1974).
There is a substantial overlap of the known distribution of the putative non-target species within the Victorian trap exemption zones where leg-hold traps and snares are used for wild dog control. In the western zone, some of the species most common to the highlands of eastern Victoria are absent (eg. superb lyrebird, spot-tailed quoll, long-nosed bandicoot and bobuck) or their distribution suggests much sparser or patchy populations overall (eg. common wombat, eastern-grey kangaroo, goanna) (Figures 2a – 2q), that may indicate a reduced potential for non-target captures. However, as distribution maps do not indicate population density, this conclusion would warrant further analysis.
5.0 Identifying Indicators Of Trapping Stress
5.1 Stress and stressors
Stress is a response to a stressor and a means to adapt to it by reducing or eliminating its effects (Webster 1998). A state of stress occurs when an animal encounters adverse physiological or emotional conditions that cause a disturbance to its normal physiological or mental equilibrium by a stressor (Manser 1992). The general adaptation syndrome (Tolosa et al. 2007) suggests that there are three generalised responses to a stressor; alarm is an initial response, followed by adaptation to the stressor that reduce or eliminate its effects, while exhaustion may result if the capacity of the animal to adapt is exceeded (Seryle 1950).
Trapping activates predictable physiological responses as a reaction to a range of stressors during capture (Moberg 1985, Kreeger et al. 1990). Ongoing stressors may have a negative impact upon the welfare of animals (Jordan 2005) and attempts to understand their impact can be made by measuring the magnitude of the biological response, pre-pathological state and consequent pathology (Moberg 1985, Carstens et al. 2000). A stressor does not lead to suffering if the animal can act without difficulty to reduce its impact, but when stressors are prolonged, too severe or multiple stressors exist, suffering can be the consequence (Webster 1998). Pathological changes and disease may result if the stressor or a combination of several stressors require the diversion of resources from other biological activities that are critical to an animal’s well being (Moberg 1985, Carstens et al. 2000). Where normal function is disrupted the potential for distress, suffering and a decline in welfare is possible (Moberg 1985, Carstens et al. 2000).
In order to make objective decisions and predictions concerning welfare states associated with trapping, the quantification of different types of stress arising from a range of stressors needs to be undertaken. Welfare science has a low level of precision when attempting to objectively measure stress, especially in a range of species, hence an assessment of an animal’s welfare often requires the use of several different approaches (Webster 1998, Dawkins 2001a). Similarly, in attempting to describe pain experienced by animals, a range of physiological as well as behavioural indicators may be needed (Rutherford 2002). The presence or absence of behavioural, autonomic or endocrine stress responses can be used as indicators of welfare states in animals. Broom (1988) lists a range of indicators used in an attempt to objectively describe an animal’s welfare; these are further summarised in four general categories:
Behavioural indicators: include indicators of pleasure and the extent to which strongly preferred behaviours can be shown. A variety of normal behaviours may be shown or suppressed or behavioral indicators of aversion (eg. avoidance) may be demonstrated;
Physiological indicators: include those that can indicate normal and abnormal physiological processes, coping mechanisms and anatomical development;
Pathological: changes such as trauma, changes in brain function, disease, immunosuppression and behavioural pathology;
Survival, growth and development: can be an indicator of welfare if it is possible to contrast normal versus reduced or abnormal life expectancy, growth or breeding.
5.2 Behavioural indicators
A wide range of common behaviours are used by animals in the expression of pain, including: escape reactions, vocalisation, aggression, withdrawing, recoiling, biting and chewing (Gregory 2005). A fearful and/or anxious domestic dogs may tuck its tail down, pin its ears against its head and display piloerection, lip licking and yawning (Neilson 2002). Vocalisation in dogs can occur due to play, excitement, communication, threat, attention seeking, defence, pain, anxiety or fear (Landsberg et al. 2003). Body posture tends to be lower with fear, anxiety or submission and common behaviours such as snout licking, body shaking, paw lifting and the amount of time that the tongue protruded were linked with increased heart rate and cortisol production in response to a stressor (Beerda et al. 1997). Certain aggressive behaviours in domestic dogs have been associated with the response to some painful stimuli (Borchelt 1983) and fear alone can release aggressive behaviours (Galac et al. 1997, King et al. 2003). In domestic dogs, the suddenness and intensity of a novel stimulus governs how effectively it will produce fear, as will a range of genetic and environmental factors (King et al. 2003). Studies of captive silver foxes showed that ear posture, activity and approach to the front of the cage could be used as indicators of welfare states, although they were not reliable in all cases (Moe et al. 2006). The absence of two behavioural indicators of poor welfare in trapped target species (self-mutilation and unresponsiveness) were used to indicate if a trap was acceptable (Harrop 2000).
An animal’s general appearance or ‘nocifensive’ behaviour is one of the few ways available to interpret its perception of pain (Carstens et al. 2000). However, it is influenced by species-specific differences in response (Valverde 2005) and applies to behaviours in response to potential tissue injury (Mersky et al. 1994). Behavioural and endocrine indicators of pain in livestock have been applied to the development of standard pain assessment in agriculture (Mellor et al. 2000, Molony et al. 2002). Pain-specific behaviours include bucking in lambs in response to wound palpation after castration, escape behaviour of calves in hot-iron dehorning and increases in high frequency calls in piglets undergoing castration (Weary et al. 2006). Acute pain escape behaviours may be modified when pain persists and guarding behaviours may be observed where an animal protects or engages in a range of strategies to protect the sensitive area (Zimmerman 1986). There are few studies of behavioural indicators of distress in marsupial fauna. Tammar wallabies (Macropus eugenii) learned to be fearful and flee a model fox and then transferred this aversion to a model cat in a set of behaviours typical of predator avoidance (Griffin et al. 2002). However, there is no comprehensive and systematic study of the behaviours of endemic wildlife species that may be used to assess their stress response to traps.
Rather than interpretation of particular behaviours and their relevance to stress, testing the strength of an animal’s motivation by measuring the sacrifice it is prepared to make to accommodate them may be an alternative approach (Dawkins 1980, Broom et al. 1993, Dawkins 1993), allowing a more objective assessment of an animal’s choice (Dawkins 2001a). Aversive learning studies may assist in understanding what stressors have caused suffering that animals wish to avoid in the future (Rushen 1996).
Post-operative pain in domestic dogs has been investigated using subjective measures such as visual analogue and numerical scale ratings, pain threshold tests (Conzemius et al. 1997), response to palpation of wounds (Pascoe et al. 1993) and other behavioural indicators such as variations in greeting behaviours to owners (Hardy et al. 1997). The accuracy of assessments of pain by scoring is limited by their subjectivity, lack of contemporary controls (ie. a comparative group that experiences ‘no pain’) and lack of positive controls (ie. a comparative group where animals are subjected to a ‘known amount’ of pain). In experiments, behavioural changes caused by some analgesics independent of pain relief are possible, as are interactions of behaviours arising from fear and apprehension associated with pain (Flecknell et al. 2004). This suggests that it may not be possible to use analgesia in experimental groups to manipulate and identify behavious that are caused by pain alone.
Some behaviours or measures can be used as correlates of animal suffering or distress that are based upon indicators such as the intensity or response to stressors. Marks et al. (2004) used activity data loggers that measured the relative duration and activity of dingoes after capture to test the effectiveness of a drug to alleviate resistance to the trap and injury. While the degree of activity cannot be used as a direct measure of distress, the degree of resistance and escape behaviour in traps is believed to correlate with the type and extent of trauma sustained (Balser 1965) and trauma is commonly scored and used to determine the welfare impact of various traps (Tullar 1984, Van Ballenberghe 1984, Olsen et al. 1988, Onderka et al. 1990, Hubert et al. 1996, Phillips et al. 1996b, Iossa et al. 2007) (see chapter 5.4). Measuring simple indicators of activity of animals in traps may be a practical way to measure relative improvement in welfare even though it cannot be used to account for the specific nature of this improvement.
5.3 Physiological indicators
Animals subjected to a stressor will release a cascade of hormones as an adaptive, short-term response to a stressor (Baxter et al. 1987). There are two main physiological stress pathways that lead to the activation of the hypothalamic-pituitary-adrenal (HPA) axis and/or the sympathetic nervous system (SNS). Corticotrophin releasing hormone stimulates the secretion of adrenocorticotrophin hormone (ACTH) from the anterior pituitary and this influences the release of glucocorticoids from the adrenal cortex that play a major role in the conversion of protein and lipids to usable carbohydrates and the breakdown of body fats. This prepares an organism to deal with a perturbation and mobilises energy stores to meet short term requirements (Korte et al. 2005). The SNS can be activated by the HPA and in general prepares an animal for ‘fight or flight’ and in doing so it causes mobilisation of glycogen and free fatty acids, increased heart rate, vasoconstriction in body regions not directly involved in fight or flight and has effects on gut motility (Gregory 2005). If the animal is unable to escape from the stressor it may adopt a mode of ‘conservationwithdrawal’ with consequent increases in pituitary-adrenocortical activity (Moberg 1985). Endogenous opioids may initially be released in response to some painful noxious stimuli with resulting stress-induced analgesia. However, more prolonged stress produces hyperalgesia which contributes to aversive and guarding behaviours (Vidal et al. 1982, Kinga et al. 2007). Any stressor may elicit an increase in circulating steroids, but in contrast to early predictions, not all stressors produce an HPA response (Mason 1968).
The measurement of cortisol has been the most commonly used indicator of stress in most mammals and non-invasive sampling methods such as salivary sampling can be used to reduce restraint artefacts (Kirschbaum et al. 1989). Restraint and venipuncture can be a significant stressor and may be a confounding factor in the measurement of stress response (Beerda et al. 1996, Hennessy et al. 1998). Values of cortisol were measured in dogs subjected to stressful situations such as loud noises (20.4 nmol/L), falling bags (18.7 nmol/L) and electric shock (15.5 nmol/L). Peak cortisol concentrations were reached shortly after the acute stimuli (between 16 to 20 minutes) and declined thereafter usually within an hour (Beerda et al. 1998). However, cortisol concentrations may not always be a good indicator of how a dog perceives prolonged exposure to a stressor, or a continuous series of stressors. Animals that are regularly subjected to stressors or have stressful lives may have enlarged adrenal glands and secrete greater amounts of cortisol (Baxter et al. 1987). Moreover, there are a range of species-specific, individual, environmental, seasonal and circadian influences on cortisol concentrations identified in canids (De Villiers et al. 1995). Comparative interpretation of cortisol concentrations as an absolute and additive measure of stress must be undertaken cautiously and in context. Nonetheless, cortisol has been used in a wide range of species to investigate stressors such as restraint, capture, transport, handling and the response to sound and predator odours (Table 3). It is important to recognise that trapping may present an array of different stressors of varying intensity throughout the duration of captivity and this places a practical limitation on how and when cortisol concentrations can be used to measure welfare outcomes (Chapter 6.1).
Table 3. Studies that used cortisol (CORT) and adrenocorticotrophic hormone (ACTH) to study stress in species responding to various stressors where the concentration were indicated as higher (H) relative to established normals, control or placebo populations.
|African wild dog||Lycaon pictus||Handling||H||De Villiers et al. 1995|
|Blue fox||Alopex lagopus||Handling||H||H||Osadchuk et al. 2001|
|Domestic dog||Canis familiaris||Transport||H||Bergeron et al. 2002|
|Domestic dog||Canis familiaris||Transport||H||Frank et al. 2006|
|Domestic dog||Canis familiaris||Acoustic||H||Gue et al. 1989|
|Domestic dog||Canis lupus||Transport||H||Kuhn et al. 1991|
|European rabbit||Oryctolagus cuniculus||Predator odour||H||Monclus et al. 2006|
|Green monkeys||Cercopithecus aethiops||Capture||H||Suleman et al. 2000|
|Grizzly bear||Ursus arctos||Capture||H||Cattet et al. 2003|
|House sparrow||Passer domesticus||Capture and handling||H||Romero et al. 2002|
|Koala||Phascolarctos cinereus||Capture||E||Hajduk et al. 1992|
|Laboratory rat||Rattus norvegicus||Predator odour||H||Thomas et al. 2006|
|Lapland longspur||Calcarius lapponicus||Capture and handling||H||Romero et al. 2002|
|Red fox||Vulpes vulpes||Trapping||H||H||Kreeger et al. 1990|
|Silver fox||Vulpes vulpes||Handling||H||Moe et al. 1997|
|Silver fox||Vulpes vulpes||Blood sampling||H||Moe et al. 1997|
|Vicuna||Vicugna vicugna||Restraint||H||Bonacic et al. 2006|
|White crowned sparrow||Zonotrichia leucophrys||Capture and handling||H||Romero et al. 2002|
There are a range of objective measurements considered to be associated with brain function during stress that have been proposed to assess welfare states. Changes in the hormone oxytocin and concentrations of neurotransmitters such as dopamine may be associated with the perception of pleasure. Event-related evoked potentials (ERPs) and the frequency spectrum of electroencephalographs (EEGs) have been found to be useful in assessing the perception of pain in humans (Bromm 1985, Chen et al. 1989) and in livestock (Barnett et al. 1996, Ong et al. 1996, Morris et al. 1997). These procedures are difficult to use in free-ranging and wild species as they cannot be used remotely and typically require surgical procedures. Increases in plasma oxytocin are associated with decreases in ACTH and glucocorticoids and proliferation of lymphocytes (Broom et al. 2004). The exposure of animals to psychological stressors or hostile environments initiates the secretion of a range of hormones that include cortisol, oxytocin, prolactin, catecholamines and renin (Van de Kar et al. 1999) and other factors such as nitric oxide (NO) modulate the immune system in response to stress (Lopez-Figueroa et al. 1998).
Measures of animal emotional responses are currently limited to a relatively simple range of physiological and behavioural responses where indicators such as stress hormones, elevation in heart rate or behaviours are attributable to fear or anxiety. These measures do not address the significance of the conscious experience, where the conscious awareness of sensations and emotions may be central to the capacity to suffer (Mendl et al. 2004). Heart rate has been used as an easily measured psychophysiological indicator of stress in dogs, yet increased heart rate may be associated with both positive and negative emotional states, and while it may be correlated with behaviours (Palestrini et al. 2005) it is difficult to use it as welfare indicator in isolation from other information to assist the interpretation of the emotional state.
In some species excitement and strenuous exercise can cause contraction of the spleen and expulsion of erythrocytes into circulation (Wintrobe 1976) that may alter normal erythrocyte numbers, haemoglobin concentrations, packed cell volume (PCV) and mean corpuscular volume (MCV) (Hajduk et al. 1992). Polymorphonuclear leucocytes include neutrophils which are the most abundant of the leucocytes. Neutrophils have the ability to migrate to the site of infection and inflammation and have a potent antimicrobial effect, but they have also been implicated in tissue damage (Schraufstatter et al. 1984, Ellard et al. 2001). Short-term mental stressors have been shown to cause a significant increase in neutrophil activation (Schraufstatter et al. 1984, Ellard et al. 2001) and this is confirmed in response to trapping stress in foxes (Kreeger et al. 1990). Neutrophil counts were significantly increased while lymphocytes decreased in dogs subsequent to air transport (Bergeron et al. 2002) and in coyotes following capture and restraint (Gates et al. 1976). Clomipramine, a tricyclic antidepressant, is used to treat anxiety disorders and aggression (Mills et al. 2002) and was supported as a treatment to mitigate transport stress in dogs as it reduced cortisol responses and neutrophil to lymphocyte (N:L) ratios compared to a placebo group (Frank et al. 2006). Neutrophil numbers increased and corresponded to an increase in the N:L ratio in koalas after capture (Hajduk et al. 1992). In vicuòa (Vicugna vicugna), animals that were restrained in enclosures showed a significant increase in N:L ratio (Bonacic et al. 2006). Similar changes in the N:L ratio have been found in pigs dosed with cortisol (Widowski et al. 1989). The injection of corticosteroids or adrenocorticotrophic hormones caused an increase in neutrophils and a decrease in lymphocytes within 2 – 4 hours in dogs (Jasper et al. 1965) and hence N:L ratios may be well associated with cortisol stress response, yet may show a delayed and flattened response. In macropods, haematological characteristics did not appear to be obvious markers of any of a range of clinical stressors including capture myopathy (Clark 2006). Variations in N:L ratios and haematological responses between species or animal groups may be unpredictable. Leukocyte counts are subject to diurnal variation, with neutrophils typically peaking in dogs during the day, corresponding to a decline in lymphocytes, which tend to peak during the mid evening (Lilliehöök 1997, Bergeron et al. 2002) and this is likely to be an important consideration if responses to less intense stressors are to be compared. In a range of studies that have sought haematological correlates with a range of stressors, N:L ratios appear to relate to capture, transport, trapping, housing and restraint stress, but appear to be less applicable to stressors that produce physical trauma (Table 4).
A range of biochemical indicators has been used to investigate a variety of stressors in different species (Table 5). Alsatian dogs that were subjected to exercise in hot temperatures showed an increase in glutamic oxalacetic transaminase (GOT), lactic dehydrogenase (LDH), phosphohexose isomerase (PHI), acid phosphatase (ACP), alkaline phosphatase (ALP), aldolase (ALD) and lipase (LIP) (Bedrak 1965). Alkaline phosphatase is found in most tissues and in high levels in bone and gut. Exercise and elevated corticosteroids can elevate ALP in dogs (Dorner et al. 1974). Conceivably, stress-induced increases in cortisols in trapped foxes could have caused the elevations of ALP (Kreeger et al. 1990). Restraint stress in mice has been shown to increase levels of LDH, creatine kinase (CK, formerly CPK), aspartate aminotransferase (AST) and alanine aminotransferase (ALT) (Sanchez et al. 2002).
Creatine kinase concentrations are used for diagnosing skeletal muscle damage in animals and exertional myopathy which is a disease of the skeletal muscles and myocardium (Aktas et al. 1993). In rats the concentration of serum CK correlated strongly with the volume of muscle traumatised by crushing injury and LDH, AST and ALT concentrations increase in response to some crushing injuries (Akimau et al. 2005). Forced or mechanical restraint will cause an elevation of CK values in human patients (Goode et al. 1977). There are three isoenzymes that predominate in the skeletal muscle (MM) and myocardium (MM and MB), and intestine and brain (BB) of dogs (reviewed in Aktas et al. 1993). Elevated CK-MM sub-fraction is typically associated with muscle trauma (rhabdomyolysis) through shock, surgery or disease affecting the skeletal muscles (Prudhomme et al. 1999). Experimental lower limb gunshot trauma in pigs also caused significant elevation of CK values (Münster et al. 2001). Shivering may induce elevation in creatine kinase (Wladis et al. 2002). Tourniquet ischemia of the arm produced with a pneumatic cuff for between 30 minutes to 80 minutes caused elevations in LDH, CK and total protein which could be detected when the cuff was applied for more than one hour and this response was detectable for three days after its removal (Rupiñski 1989).
Table 4. Major haematological values measured in various species in response to stressors and their concentration as higher (H) or lower (L) relative to established normals, control or placebo groups (N:L = neutrophil to lymphocyte ratio, WCC = white cell count, RBC = red blood cell count, Hb = haemoglobin, PCV = packed cell volume, GRA = granulocytes, LYM = lymphocytes, EOS = eosinophils, NEU = neutrophils)
|Domestic dog||Canis familiaris||Transport||H||L||H||Bergeron et al. 2002|
|Domestic dog||Canis familiaris||Transport||H||H||Frank et al. 2003|
|Domestic dog||Canis familiaris||Transport||H||H||H||Kuhn et al. 1991|
|Eurasian otter||Lutra lutra||Capture||H||L||L||H||Fernandez-Moran et al. 2004|
|Flying fox||Pteropus hypo-melanus||Restraint||L||L||Heard et al. 1998|
|Grizzly bear||Ursus arctos||Capture||H||L||L||H||Cattet et al. 2003|
|Human||Homo sapian||Mental stress||H||H||Ellard et al. 2001|
|Kit fox||Vulpes macrotis mutica||Capture||H||L||McCue et al. 1987|
|Koala||Phasco-larctos cinereus||Capture||H||H||H||H||H||L||H||Hajduk et al. 1992|
|Laboratory dogs||Canis familiaris||Housing||H||H||H||Spangenburg et al. 2006|
|Red fox||Vulpes vulpes||Trapping||H||H||H||Kreeger et al. 1990|
|River otter||Lontra canad-ensis||Capture||H||Kimber et al. 2005|
|Silver fox||Vulpes vulpes||Handling||L||L||L||Moe et al. 1997|
|Silver fox||Vulpes vulpes||Blood sampling||L||L||Moe et al. 1997|
|Vicuna||Vicugna vicugna||Restraint||H||H||Bonacic et al. 2006|
Table 5. Major blood biochemistry values measured in various species in response to stressors and their concentration as higher (H) or lower (L) relative to established normals, control or placebo populations (BR = bilirubin, UR = urea, Na = sodium, Gl = glucose, Ca = calcium, Glob. = globulin, Cr = creatinine, K = potassium, LDH = lactate dehydrogenase, CK = creatine kinase, AST = aspartate aminotransferase, ALT = alanine aminotransferase, ALP = alkaline phosphatase, Chl. = cholesterol).
The use of CK is a specific marker for diagnosis of muscle disease (0.83 specificity) (Aktas et al. 1993), however its reliability is influenced by snake venom toxicosis, myocardial disease associated with parvovirus, dirofilariasis, haemolysis and venipuncture that penetrates muscle tissue and some therapeutic agents (reviewed in Aktas et al. 1993). In flying foxes (Pteropus hypomelanus) short-term restraint was associated with changes in haematology and blood biochemistry which were significantly reduced by anaesthesia with isoflurane (an anaesthetic) (Heard et al. 1998). The progressive evaluation of recently captured river otters (Lontra canadensis) showed that CK and AST/ALT were not good indicators of musculoskeletal injury owing to probable interactions with existing pathology due to infection, parasitism and other factors independent of capture injury (Kimber et al. 2005). Similarly, elevation of ALT in dogs has been shown to be associated with skeletal muscle degradation and not liver damage (Valentine et al. 1990). Some stressors may not be detected in some species or breeds given variation in response, genotypic difference or different context. For instance, in Alaskan sled dogs after long distance races there is little indication of increases in serum CK values as an indication of skeletal muscle damage after days of strenuous racing (Hinchcliff 1996), yet elevation of CK is associated with physical exertion in other domestic dogs (Aktas et al. 1993).
Overall, the most commonly used biochemical indicators of stress associated with capture, handling, injury and transport are CK, AST, ALT and ALP and changes in the values of these indicators have been successfully used to reveal stress responses in a wide range of animals (Table 5).
5.4 Visible pathological indicators
One criterion for the assessment of the humaneness of leg-hold traps has used the incidence and extent of physical injury as the primary indicator of trap welfare outcomes. In most studies the assumption is that the extent of trauma is inversely proportional to the relative humaneness of the device. Trauma scales have been used to assess injury produced by various traps and snares (Tullar 1984, Van Ballenberghe 1984, Olsen et al. 1988, Onderka et al. 1990, Hubert et al. 1996, Phillips et al. 1996a) and are reviewed by Iossa et al. (2007). Many studies have used damage scores based upon the extent of the visible trauma inflicted upon the captured limb only (eg. Olsen et al. 1986, Houben et al. 1993, Fleming 1998, Stevens and Brown 1987). Whole body necropsies attempt to fully account for the entire range of injuries, such as puncture wounds caused by vegetation (Hubert et al. 1997) that occur during trapping. Some authors have ignored mouth injuries (eg. chipped and broken teeth, lacerations and abrasions of the gums and lips), yet these are common injuries in carnivores caused by traps (Onderka et al. 1990). Bite wounding (Marks et al. 2004), predation and death of animals held in a trap have not always been regarded as relevant to the welfare outcomes and performance of particular devices (eg. Fleming et al. 1998). Many of the earlier scoring systems did not account for injury and debilitation associated with pathology such as capture myopathy (Tullar 1984, Van Ballenberghe 1984, Olsen et al. 1988, Onderka et al. 1990) and given the various manifestations and progression of this disease (Chapter 6.2.3), it is likely that gross observations would be inadequate to diagnose this condition.
Since the development of injury scoring, there has been an increase in the number of injury classes used in various studies (Onderka et al. 1990, Phillips et al. 1996a, Hubert et al. 1997) and altered weighting and scoring methods make comparisons between many studies difficult (Engeman 1997, Shivik et al. 2000, Iossa et al. 2007). Van Ballenberghe (1984) developed five classes of injury scores to assess trap injury (Table 6).
Table 6. Trap injury classification system developed by Van Ballenberghe (1984).
|I||Slight foot/leg oedema, no lacerations or broken bones.|
|II||Moderate oedema, lacerations less than 2.5 cm long, no broken bones and joints.|
|III||Lacerations at least 2.5 cm long, visible tissue damage, no tendon damage, one metacarpal or phalanx bone broken.|
|IV||Combinations of deep, wide lacerations, severed tendons, broken metacarpals, broken radius or ulna bones and joint dislocations.|
Stevens and Brown (1986) developed a rating system that was based upon that by Van Ballenberghe (1984) in order to investigate the humaneness of steel-jawed traps and treadle-snares to captive target and non-target vertebrates in Victoria. These authors modified some of the classification and added an additional one to assist in discerning between slight injuries and total absence of injury (Table 7).
Table 7. Trap injury classification system used by Stevens and Brown (1987) based upon that developed by (Van Ballenberghe 1984).
|I||No visible trap-related injuries.|
|IV||Bones disjointed or broken.|
Using the criteria of Stevens and Brown (1987), Murphy et al. (1990) constructed two broad classifications of "major injury" which included ratings 4, 5 and 6 and "minor injury" for any of ratings 1, 2 and 3. It was assumed that animals with minor injuries would not be permanently debilitated upon release. Meek et al. (1995) and Fleming et al. (1998) used a scoring system following that of Van Ballenberghe (1984) in an Australia-wide analysis of trauma caused by a range of traps (Table 8).
Table 8. Injury classes attributed to target and non-target animals (after Van Ballenberghe 1984) with inclusion of Class V (Fleming et al. 1998).
|I||No visible trap-related injuries or only slight foot and /or leg oedema with no lacerations and no evidence of broken bones or dislocated joints.|
|II||Moderate oedema with skin lacerations 2.5 cm or less, bones and joints as in Class I.|
|III||Skin lacerations greater than 2.5 cm long with visible damage to the underlying tissue, tendons intact, bone breakage limited to one phalanx or metacarpal / tarsal.|
|IV||Various combinations of deep and wide lacerations, severed tendons, broken metacarpal/tarsal, radius, tibia, fibula and ulna bones, joint dislocation of the legs, and/or amputation.|
|V||Dead in trap from hyperthermia/hypothermia, excessive blood loss, shock or capture myopathy.|
The first widely used scoring systems for trapping trauma (Table 9) sought to weigh individual injuries in terms of their potential to cause incapacitation and impact upon the welfare of animals (Onderka et al. 1990). These systems were additive and allowed quantification and comparison of mean or median injury scoring developed for different devices. Given that they accommodated a wide range of specific injuries the resolution of this approach was greater and allowed researchers a greater ability to detect differences in injury outcomes from a range of devices.
Table 9. Trauma scoring system adopted by Onderka et al. (1990).
|1-5||Edematous swelling and/or hemorrhage.|
|1-5||Cutaneous laceration <2 cm long.|
|10||Cutaneous laceration >2 cm long.|
|10-20||Subcutaneous muscle laceration or maceration.|
|20-40||Tendon or ligament maceration with partial severance.|
|30||Partial fracture of metacarpi or metatarsi.|
|30-40||Fracture of digits.|
|30-40||Amputation of digits.|
|50||Joint luxation of digits.|
|50||Simple fracture below carpus or tarsus.|
|50||Severance of tendons below carpus or tarsus.|
|75||Compound fracture below carpus or tarsus.|
|100||Simple fracture above carpus or tarsus.|
|200||Compound fracture above carpus or tarsus.|
|200-300||Luxated elbow or hock joint.|
|400||Amputation of limb.|
Table 10. Trauma scoring system (summarised) adopted by the International Organisation for Standardisation (ISO) and subject to threshold assessments (Harrop 2000).
|5||Minor cutaneous laceration.|
|10||Major cutaneous laceration.|
|25||Severance of minor tendon or ligament.|
|25||Amputation of one digit.|
|30||Permanent tooth fracture exposing pulp cavity.|
|30||Simple rib fracture.|
|50||Amputation of two digits.|
|100||Amputation of three or more digits.|
|100||Spinal chord injury.|
|100||Compound rib fractures.|
|100||Ocular injury resulting in blindness in an eye.|
The ISO committee restricted definition of welfare impacts associated with trapping to purely pathological observations (Harrop 2000). Their trauma scores permitted an agreed level of trauma to be associated with an ‘unacceptable’ trap that would allow major debilitative injury in a majority of cases (Table 10). It was established that the acceptability of a trap would be contingent upon a 90% confidence that it would exceed a lower threshold score on 50% of occasions and an upper score for 20% of occasions (Harrop 2000).
Poor welfare outcomes arising from trapping have been defined by the existence of pathological signs that increase in their severity from the lowest to highest score (Anon 1997, Shivik et al. 2005) and this approach was useful in relating trauma that would be considered to be an unacceptable welfare outcome:
- Fractures and/or joint luxation proximal to the carpus or tarsus;
- Severance of a tendon or ligament;
- Major periosteal abrasion;
- Severe external haemorrhage into an internal cavity;
- Major skeletal muscle degeneration;
- Limb ischemia;
- Fracture of a permanent tooth exposing the pulp cavity;
- Ocular damage including corneal laceration;
- Spinal cord injury;
- Severe internal organ damage;
- Myocardial degeneration
5.5 Survival, growth and development
Trapping studies usually assume that the probability of capture for all individuals in a population is equal. Trap related injuries and debilitation can reduce the chances of recovery from subsequent trapping (Earle et al. 2003). This permits trap-release-recapture studies to provide some insight into the relative impact of traps upon a population. Studies that use radio-collars to monitor the long-term fate of animals subsequent to capture and release are probably the most informative in allowing the fate of animals to be known. After trapping a population of Rüppels fox (Vulpes rueppellii) using padded leg-hold traps, the majority of individuals were given low injury scores but survival was reduced possibly due to higher levels of predation upon foxes as trapping could have caused limping or debilitation in other ways (Seddon et al. 1999, in Iossa et al. 2007). There is an absence of studies that record the survival, growth and development of animals after capture as a primary aim of the study. Such controlled studies are difficult to conduct, as good experimental design would ideally require a population of animals that have not been trapped to be similarly monitored.
5.6 Discussion and conclusions
The value of behavioural indicators of stress is probably limited in the comparative assessment of leg-hold traps. As behaviours can be variable and not specifically related to stress, they can be readily misinterpreted (Beerda et al. 2000). Wild animals may hide symptoms of pain and distress that might otherwise make them vulnerable to predators and they may display different signs and symptoms of pain (Jordan 2005). Moreover, the utility of human experience and direct observation to infer the suffering of animals is limited, as we often do not have the same perceptual abilities; eg. the absence of a vomeronasal organ; inability to detect infrared radiation, magnetic fields, specific pheromones and some sound frequency ranges (Gregory 2005).
Studies of the aversiveness of different trapping devices require ‘choices’ to be made between different traps. Trapping activities are not undertaken in a manner where a wild animal can be easily observed or be given multiple exposures to traps, and animals are usually unaware of the trap prior to capture. Further, it is likely that all trapping stressors are intense regardless of the device used. Discerning discrete differences in behaviour or preference could be difficult to interpret and may provide very limited information about welfare states. While the long-term control of coyote populations was found to reduce trapping success and may have suggested individual aversion (Sacks et al. 1999), problematically this may have been the consequence of selected neophobia at a population level, rather than individual preference based on prior experience. Repeated cage trapping did not affect the recapture rate of red foxes (Baker et al. 2001), yet other studies using endocrine, pathological, biochemical and haematological indicators suggest similar traps produce significant stress in foxes (White et al. 1991). It is possible that cage trapping did not produce recognisable aversive behaviour for a range of reasons, such as prior negative states not being intense enough to promote learning, rapid extinction of the memory of prior captures, and/or failure to differentiate the cage trap from other features in the environment upon recurrent capture. It is difficult to support an assumption that a lack of demonstrated learned aversive behaviour (perhaps from one or few experiences) equates to an overall lack of aversiveness and absence of trapping stress. Moreover, behavioural indicators of trapping stress may not provide sufficient sensitivity to discern between subtle differences in welfare outcomes from different trap devices.
Trauma scales and scores are limited in their ability to assess the overall welfare impact of trapping. The nature of suffering associated with injuries; long-term impacts of injury upon survival, resulting changes in fecundity and impacts upon dependent animals cannot be known (Iossa et al. 2007). Variation and lack of compatibility in various trauma scales and their application even within one species makes comparison of studies difficult (Engeman 1997). One key deficiency is that the amount of time that an animal spends in captivity is rarely known even with moderate accuracy. Monitoring trapping practices in the field must contend with a wide range of experimental variables such as heterogeneous habitat and age structure of the population; differences in light, temperature and precipitation; trapping protocols and variations in the performance of each trapping device. Under such conditions, subtle changes in welfare states may be difficult to detect. Trap injury categories and scoring systems may be capable of discerning differences in large magnitudes of gross physical injury associated with a range of traps, especially if they are tested contemporaneously, yet severe injury is an endpoint of poor welfare. Improvements in trapping practices that may incrementally improve a range of welfare outcomes may be difficult to demonstrate given the problems in controlling experiments, the high degree of experimental variance in such field assessments, and the fact that trauma is only one component of trapping stress that is relevant to assessing welfare impacts.
An ISO Technical Working Group rejected the adoption of hormone and blood analysis as indicators of welfare, although this was an approach favoured by European scientists (Harrop 2000). Monitoring the neuroendocrine systems is difficult to do without introducing stressors associated with blood sampling and restraint normally associated with such investigations, which may confound experimental results (Carstens et al. 2000). However, other physiological indicators such as N:L ratio, ALP, AST, ALT and CK appear to be useful indicators of stress that have been consistently associated with known stressors and pathology in a wide range of species. Creatine kinase appears to be one of the most useful indicators of trapping stress, given its potential sensitivity to skeletal muscle trauma, exertion and myopathy, which are key poor welfare outcomes. Stress leukograms may be useful if used appropriately in experiments to assess stress, and N:L ratios in particular have been used to assess a wide range of stressors (Marks, in review, Appendix 1). Unlike collection and measurement of cortisol, these indicators are comparatively slow to respond and are less likely to be affected by blood sampling and handling stress, although N:L ratios and ALP levels are probably correlated with the release of cortisol. A significant drawback is that not all species will respond to stress indicators in a uniform way and the most appropriate use of haematological and blood biochemistry indicators will depend upon an understanding of these species differences. Using data logger systems that reveal the capture period and relative activity of animals in conjunction with physiological indicators such as CK, AST, ALP, ALT and N:L ratios and detailed whole-body necropsies is likely to yield the most useful, practical and unequivocal insight into the relative welfare impacts of traps. Many of the haematological and biochemical indicators are standardised, cost-effective and widely available laboratory tests that, if properly applied, could provide sufficient information to monitor relative welfare states and promote adaptive management of trapping practices towards better welfare outcomes.
6.0 Stressors And Pathology Associated With Leg-Hold Trapping
The stressors produced by trapping and the resulting stress and pathology that may arise is directly related to the potential negative welfare outcomes associated with trapping and snaring. Improving the welfare outcomes of trapping will require the removal or reduction in the intensity of various stressors. As species respond to various stressors in different ways, the contribution of each stressor towards the welfare state of each species should be considered independently. A model of an animal’s response to stressors suggests that exposure to stressors can overwhelm an animal’s defence of its normal biological functions and result in prepathological or pathological states (Figure 3).
Figure 3. Model of the response of trapped animals to stressors and stress associated with leg-hold traps (Modified from Moberg 1999 and Carstens and Moberg 2000).
6.1 Trapping stressors
Startle response (often referred to as fright) occurs when an animal encounters a perceived danger without being prepared for it (Gregory 2005). Exposure of an animal to novelty is one of the most potent conditions that can lead to a negative emotional response (Dantzer et al. 1983, King et al. 2003) and the fearfulness that it produces will be influenced by the physical characteristics of its presentation, including its proximity, intensity, duration and how suddenly it appears (Russel 1979). Perception of sudden movement is believed to a potent stressor in provoking fear in domestic dogs, but its extent depends upon the nature of the stimulus (King et al. 2003). Leg-hold traps and most snares are hidden, and activation of some traps occurs within 18.52 -18.59 ms (Johnston et al. 1986) and correspond to velocities of between 5.38 - 6.83 m s-1 with an impact forces of 182.3 - 281 N in Victor Soft-Catch traps (Earle et al. 2003). The suddenness and forcefulness of the initial activation and restraint by a leg-hold trap is highly likely to be a potent cause of startle response.
6.1.2 Primary trauma and acute pain
Primary trauma caused by trapping occurs immediately upon capture or quickly thereafter. Trauma is defined as tissue injury that usually occurs suddenly as a result of a violent action that is responsible for the initiation of the HPA, metabolic and immunological responses (Muir 2006). Such events will usually generate pain, stress and fear. Collectively, these reactions normally benefit animals by enabling them to avoid situations that cause trauma and will prevent further injury and compensate to restore homeostatic function (Foex 1999). Pain is usually defined as an unpleasant subjective physical and emotional sensation (Bateson 1991). The International Association for the Study of Pain defines pain as “an unpleasant sensory and emotional experience associated with potential or actual tissue damage” (Mersky 1994). Pain has a unique status in that it is probably best thought of as a stressor as well as a form of stress associated with trauma in a feedback mechanism. Pain may exacerbate struggling and other behaviours that result in further trauma and pain.
Acute pain associated with trapping may be associated with primary trauma due to capture. ‘Nociceptors’ are nerve fibres specialised in the reception and transmission of noxious stimuli that elicit the release of neurotransmitters. They are located in the skin, viscera, muscles, fascia, vessels and joint capsules and respond to mechanical, thermal or chemical stimuli (Covington 2000). Pain signalling has been described as operating in several modes: control state (normal); suppressed; sensitised; or reorganised (pathologic) (Woolf 1994). Accordingly, pain is not a simple ‘hard-wired’ response that is experienced predictably and uniformly over time or between individuals. The perception of pain is modulated and attenuated by a wide range of physiological mechanisms that can enhance or reduce the experience of pain (Covington 2000).
A putative list of tissues that differ in their sensitivity to pain arranged from most to least sensitive include: cornea, dental pulp, testicles, nerves, spinal marrow, skin, serous membranes, periosteum and blood vessels, viscera, joints, bones and encephalic tissue (Baumans et al. 1994, Martini et al. 2000, Rutherford 2002). Pain from broken bones arises from distortion and pressure on receptors serving the intramedullary nerve fibres; stretching of the receptors in the periosteum, receptors in the muscle and soft tissue around the bone. The pressure resulting from haematoma triggers further pain from the soft tissues and bradykinin, histamine, potassium and peptide neurotransmitters accumulate and sensitise nociceptors and initiate tenderness (Gregory 2005). Some forms of environmental and physiological stress can modulate pain perception and are often referred to as “stress-induced analgesia” (Amit et al. 1986) and this is well documented in cases of severe trauma in humans who may not report pain for minutes to hours subsequent to injury, and is common with injuries such as fractures (5 min), cuts (13 min) and lacerations (21 min) (Melzack et al. 1982). However, the relevance or extent to which stress-induced-analgesia has a role in the mitigation of pain associated with trapping injury is so poorly known that it could not be relied upon to de-emphasise the likelihood that pain is a consistent outcome of trauma caused by leg-hold traps in all vertebrates.
6.1.3 Restraint and handling
Where species are unable to control or escape from a stressor or trauma, they may show enhanced emotional stress and responses (Seligman 1972). Restraint is one of the most common stressors experienced by animals and is a problem associated with a wide range of agricultural practices where animals are handled (Gregory 2005). Large flight distances and extreme wariness of humans is a common characteristic of wild mammals that have not been tamed or domesticated (Price 1984), and handling to euthanase or release animals from a trap can be a major stressor. Trap escapes have been recorded when traps that have held red foxes for many hours are approached by trappers, and this may suggest that greater struggling is produced by this stressor (C.A. Marks, personal observations). It is possible that the additional motivation that intensified escape behaviour is associated with a fear of humans that might produce significantly greater motivation than the combined stressors encountered prior to human contact. Selective breeding of confidence traits produces a reduction in stress associated with human contact in foxes and other species (Kenttämines et al. 2002, Trut et al. 2004). This implies that in the absence of domestication to reduce stress associated with innate avoidance of humans (Price 1984), approach and handling of wild species by humans can be expected to be more stressful than for domestic animals.
6.1.4 Behavioural, social and spatial dislocation
‘Behavioural needs’ are activities that an animal is compelled to perform such that its welfare is diminished when it is deterred from doing so (Friend 1999). Behavioural deprivation is often referred to in terms of the denial of behavioural needs (Morgan et al. 2007). The spatial requirements of an animal are normally determined by a range of factors such as the need to seek food and water, social interactions, shelter and other resources, and the home range used may vary due to season, status or other requirements in the pursuit of these needs (Price 1984). Captivity prevents the pursuit of normal behavioural needs, social interactions with con-specifics and patterns of established range and use of resources.
6.1.5 Loss of cover
Shelter and hiding is a common defensive behaviour for concealment and protection (Blanchard et al. 2001) and as a means of escape from predators and aggressive social partners (Price 1984). Trapping and restraint of animals reduces the ability of animals to retreat to cover in response to the stress it produces. Prepared diurnal shelter sites may be used during the day for nocturnal species such as foxes (Marks et al. 2006) and wombats (McIlroy 1977). Larger macropods such as eastern grey kangaroos and red-necked wallabies select open shelter sites to ensure early detection of predators in order to promote escape (Jarman 1991, Le Mar et al. 2005) while smaller macropods such as swamp wallabies rely upon cryptic shelter places to avoid detection and predation during the day (Jarman 1991, Le Mar et al. 2005). A wide range of anti-predator behavioural adaptations have evolved (Kavaliers et al. 2001) and animals respond to avoid aversive stimuli using a narrow array of ‘species-specific defence reactions’ (Bolles 1970). For instance, the defensive reactions of wild species that receive an unexpected shock from an electric fence are predominantly flight or withdrawal to a prepared retreat (McKillop et al. 1988). Common wombats were observed to immediately retreat towards their burrow system in response to aversive stimuli such as shock from an electric fence (Marks 1998a) as do swamp wallabies (C.A. Marks, unpublished data) and kangaroos (McCutchan 1980). Trapping stressors are likely to trigger species-specific defence reactions in a wide range of animals, but restraint prevents the performance of behaviours that are typically used to mitigate such stressors.
Circadian rhythms adopted by animals use light as the primary source of temporal information that is often the key cue for tightly regulated physiological and behavioural functions (Mohawk et al. 2005). The exposure of captured animals to abnormal light intensity or the disruption of their usual dial rhythms is an important stressor. Light intensity influences the activity patterns of a range of carnivores in a species-specific manner. While red foxes were found to be nocturnal 90% of the time, other carnivores are most active during the day and increased light intensity can either inhibit or promote activity (Kavanau et al. 1975). Dingoes in the NSW highlands were found to be active throughout the day, with activity peaks at dawn and dusk (Harden 1985), but in SW Queensland capture times appeared to suggest predominant nocturnal activity (Marks et al. 2004) (see Chapter 8.4). Increasing light intensity when rats lack cover increases the level of threat (Tachibana 1982) and stress can alter the use of phototic regulation for their circadian rhythms (Mohawk et al. 2005). The increase of startle response in rats tested in bright light has an evolutionary basis as rats are generally nocturnal and are more vulnerable to predators in the light (Walker et al. 2002). Nocturnal or diurnal habits of species can be typically identified by the characteristics of the photoreceptors in their retina and the predominance of rods, while diurnal species typically have higher densities of cones (Peichl 2005). Most Australian mammals are nocturnal in habit and seek shelter during daylight hours; the numbat (Myrmecobius fasciatus) is the only truly diurnal marsupial (Strahan 1984). Nocturnal activity is argued to be a primary anti-predator mechanism for many arboreal species (Goldingay 1984). Brushtail possums alter the intensity of their foraging activity in response to moonlight (Coulson 1996) and most birds (with the exception of owls, nightjars, night herons etc) are diurnal species that roost or shelter during the evening (Schodde et al. 1990). Recommendations for live trapping of nocturnal animals require that traps are set before dusk and inspected as soon as possible after dawn in order to reduce stress associated with subjecting nocturnal animals to direct light (Sharp et al. 2005a; 2005b, Anon 2007). Trapping protocols that extend the period between trap setting and daytime inspection can be assumed to increase the significance of light exposure as a stressor.
Sounds may be a powerful stressor for captive animals that cannot retreat from them. Sound stressors associated with predators can cause stress and myocardial necrosis due to insufficient perfusion of the heart muscle that can lead to death. This has been demonstrated to occur in species such as ground squirrels and rats that were made to listen to recordings of cat-rat fighting (Gregory 2005). Traps have a wide range of moving parts with attachments, chains and mechanisms that produce a varying amount of sound when activated and resisted by captive animals. Loud noises were shown to be aversive to domestic dogs and affected gastric motility and hormone release (Gue et al. 1989), activity and behaviour (King et al. 2003). Noise is an important stressor that affects the welfare of captive laboratory animals (Jain et al. 2003). In a forest habitat, ambient noise levels ranged from 40 – 70 dB while in savannah habitats it was 20 – 36 dB (Waser et al. 1986). The sound of metal on metal during cage cleaning in a laboratory was measured to be 80 dB and had a wide spectrum of harmonics that were rich in different frequencies (Morgan et al. 2007). Noise made by the capture device may compound stress experienced by the captured animal and contribute to the initial startle responses. When inspecting fox trap lines that also used Victor Soft-Catch #3 traps, treadle-snares holding foxes were heard up to 50 m away by a characteristic ‘metal against metal’ sound of the treadle plate, the chain moving through the eye of the main spring and the sound of the device hitting hard surfaces. In contrast, Victor Soft-Catch #3 traps appeared to make far less sound if they were tethered on a short chain and fox captures could not be heard until a close approach was made to the trap site (C.A. Marks, personal observations). Post-capture noise could be hypothesised as a possible contributing reason why comparative blood biochemistry values for foxes trapped in treadle-snares and Victor Soft-Catch traps differed significantly (Marks, in review, Appendix 1).
6.1.8 Food and water
No organism has a uniformly available food source and periods of negative energy balance will be normally encountered (Millar et al. 1990). Long-term captivity and restraint will not allow animals to pursue their normal foraging activities in order to meet metabolic requirements that may be exacerbated by trapping stress and mobilisation and use of energy stores. The inability to use behavioural strategies to avoid heat loss may further produce a negative energy balance. Confinement by a trap device is likely to produce a degree of food and water stress, depending upon the duration, environmental conditions, activity of the trapped animal and its nutrition and hydration upon capture. In many terrestrial vertebrates, the majority of fluids are ingested as part of the food they consume and an inability to forage for food will compromise hydration and induce thirst (Gregory 2005). In dogs, evaporative loss from cutaneous surfaces or by panting, salivation or urination (Ramsay et al. 1991) may be influenced by temperature and stressors. In laboratory conditions, at room temperature after radiant heating raised the dorsal skin temperature up to 45oC, evaporative loss was the equivalent of running 7-10 km h-1 . When dogs ran under heat their water loss increased to 85150 g hr-1 (O'Connor 1977). Dogs tend to drink water voluntarily once water loss is ≈ 0.6% of body mass (O'Connor 1977). In a hot and exposed environment, it is likely that water loss during a period of many hours resisting a trap will be significant.
The field metabolic rate (FMR) and water turnover of various animals has been calculated using a range of methods including a ‘doubly labelled’ water method (Nagy 2005). The relationship between the body mass of various vertebrate groups and FMR has been investigated by allometric scaling to describe their energetics (Nagy 2005), although the precise relationship between body mass and energy metabolism is a complex multivariate relationship (Heusner 1985). In NSW, the influx of water for adult foxes was found to be 577 mL day-1 and 444 mL day-1 for males and females respectively in November and decreased to a mean of 314 mL day-1 and 251 mL day-1 for males and females in April. Higher water intake in November may have been due to supplementation of water by drinking (Winstanley et al. 2003). As foxes obtain most of their water requirement from prey, a water intake of 314 mL day-1 corresponds to 370 g of mammalian prey ingested per day (Saunders et al. 1993). Common wombats were found to require 694 g day-1 and 1450 g day-1 of dry matter to meet their energy requirements in the dry and growing seasons respectively (Evans et al. 2003). Birds have to relatively use approximately 20 times more energy each day to live in contrast to a lizard, while mammals require 12 times more (Nagy et al. 1999). While it has been assumed that animals increase their energy expenditure in winter to meet the higher cost of thermoregulation, this has not been supported by studies that suggest that seasonal variations in metabolic rate is marginal (Nagy et al. 1999). However, given that the stress of capture will have significant metabolic costs, trapping stress is likely to be high and compounded by a need to defend body temperature if exposed to unfavourable climatic conditions.
Avoidance of predators by their detection by odour (olfaction) is a commonly used strategy in animals. Captured animals are unable to escape or avoid odour stressors. Components of fox urine have been shown to elicit endocrine and stress responses in rodents (Soares et al. 2003) and predator odours can in general produce powerful avoidance behaviour (McGregor et al. 2002). Rats avoided ferret odours and developed a sensitised stress response after the first exposure (Masini et al. 2006) and mongoose (Herpestes auropunctatus) odour was found to be repellent to rats (Rattus sp.) (Tobin et al. 1995). The sensitivity of canine scent identification is well recognised, as is their ability to detect and discern human scents at low concentrations (Lorenzo et al. 2003), and it is likely that odour detection will be a significant stressor associated with detection, avoidance, fear and anxiety associated with interactions with humans.
Some animals are strongly dependent upon behavioural thermoregulation to regulate their body temperature (Brown 1984, Brice et al. 2002) and nocturnal activity rhythms are common in order to minimise water loss and avoid high temperatures. The denial of shelter through trapping and captivity and alteration of normal activity rhythms that assist behavioural thermoregulation may cause thermal stress in unfavourable environments. The capacity for trapping to expose animals to thermal stressors will be largely dependent upon the climate, degree of shelter, season, period of captivity and species-specific attributes that determine susceptibility to thermal stress.
6.2 Trapping pathology
6.2.1 Secondary trauma and pain
Post-capture activity and secondary trauma
Balser (1965) observed that injuries caused to coyotes by steel-jawed traps were largely produced by their struggle to escape the trap and chewing of the restrained appendage. Van Ballenberghe (1984) noted that 41% of 109 wolves captured in leg-hold traps incurred severe injuries to their feet and legs. Injury sustained by wolves was thought to be directly related to the degree of struggling after capture (Frame et al. 2007). The ‘aggressiveness’ of coyotes measured by their degree of vocalisation and lunging on removal from neck snares was positively related to the degree of injury that they had sustained (Pruss et al. 2002). Much of the trauma produced from trapping is unlikely to be visible immediately or even within some hours of capture and may take many days to develop into recognisable pathology. The relationship between initial trauma and the development of secondary trauma is unclear, yet may include a wide range of physical injuries that have been documented to be caused by different trapping devices including: oedematous swelling; haemorrhage; lacerations or maceration of skin and muscle; laceration, maceration or severance of tendons and ligaments; fracture of metacarpi, metatarsi, digits and other bones; luxation of joints; compound fractures and amputation (Van Ballenberghe 1984, Linhart et al. 1986, Olsen et al. 1986, Linhart et al. 1988, Olsen et al. 1988, Fleming et al. 1998, Pruss et al. 2002, Frame and Meier 2007).
In red foxes trapped in padded and unpadded leg-hold traps, physical activity due to struggling was intense following capture, but decreased rapidly during the first two hours of capture after which struggling was intermittent (Kreeger et al. 1990). In box traps, foxes were found to be active for 35.7 ± 8.8 (SE) % of the time overall, although activity was most intense immediately after capture (White et al. 1991). A similar activity pattern was observed in dingoes that had been captured in padded Victor Soft-Catch traps, where activity was most intense during the first hour of capture, yet progressively declined to half the value in the second hour and almost a quarter by the fifth hour of captivity. Dingoes that had been captured with a trap fitted with a tranquilliser trap device (TTD) containing diazepam had significantly lower activity, especially from the second hour of capture, corresponding to the onset of the sedative/anxiolytic used (Figure 4) (Marks et al. 2004).
As bone strength increases during maturation until approximately 30 weeks of age in domestic dogs (Jonsson et al. 1984), the bones of young canids may be more susceptible to breakage. Most studies identify the swelling of the foot to be associated with foot-snares and traps, yet tend not to indicate that this is a serious injury (Logan et al. 1999, Frank et al. 2003, Iossa et al. 2007). Trap injury scoring that focuses only upon the limbs of trapped coyotes was found to be 15% lower than injuries scored when the entire body was necropsied (Hubert et al. 1997) and this suggests the need for whole body examination of trapped animals (Onderka et al. 1990) as surrounding vegetation can cause entanglement, trauma and puncture wounding (Logan et al. 1999, Powell 2005) and other trauma independent of the trap.
Figure 4. Mean hourly activity measure, AUC (Area Under Curve), for dingoes captured in Victor Soft-Catch #3 traps with a tranquilliser trap device (TTD – grey shading) (n = 19) or a placebo TTD (open bars) (n = 20) (P < 0.05) (after Marks et al. 2004).
The welfare implications of dental injures that expose the pulp cavity are highly significant as this is proposed to be the second most sensitive tissue that can produce intense pain (Baumans et al. 1994, Martini et al. 2000, Rutherford 2002). Van Ballenberghe (1984) observed that 44% of 109 wolves captured in steel-jawed traps had serious foot injuries, while 46% broke teeth. Broken, chipped or dislodged teeth occurred in 44% of adults (n=202) and 14% of juveniles (n=104) captured in steel-jawed traps (Kuehn et al. 1986). Mouth-injuries, such as chipped and broken teeth, and lacerations and abrasions of the gums and lips occur as a result of the trapped animal biting at the trap and are more prevalent in carnivores. The biting of traps is believed to be a common initial response to capture in wolves (Sahr et al. 2000) and is common in dingoes captured with Victor Soft-Catch #3 traps due to the chewing and biting of traps, probably in the initial period after capture (Marks et al. 2004). This can result in damage to metal parts of the trap (Figure 5). Englund (1982) found that 19% of juvenile foxes captured in Victor #2 and #3 long-spring traps suffered severe dental injuries by chewing traps, while 58% of adults were affected. Severe dental injury and jaw breakage may render animals unable to continue with a normal diet that requires a ‘killing bite’. The predation of livestock by larger carnivores may be associated with a need to seek alternative food after damage to dentition and such infirmity was proposed to account for lion attacks upon humans (Patterson et al. 2003) and jaguar (Panthera onca) attacks upon livestock (Rabinowitz 1986). However, injuries are also observed in the mouths of untrapped animals and some authors ignored the assessment of tooth damage due to difficulties in determining if these injuries were related to trapping alone (eg. Fleming et al. 1998).
Figure 5. Damage to ‘Paws-I-Trip®’ pan tension device ‘dogs’ caused by chewing (indicated by arrows) of the Victor Soft-Catch #3 traps after capture of dingoes (after Marks et al. 2004).
Predation and insect attack
The predation and death of non-target animals trapped in leg-hold traps is well documented and the confinement of individuals in leg-hold traps is a major disadvantage to animals that may need to defend themselves against aggressive interactions with competitors, predators or insects. Bite wounding among domestic dogs is a well recognised cause of trauma that results in severe bruising and crushing injuries (Shamir et al. 2002). A fresh bite wound to the scrotum of a trapped dingo was apparently inflicted by a con-specific (Marks et al. 2004). The predation of non-target species by wild dogs or foxes while they are held in leg-hold traps and snares has been reported in Australia (Bubela et al. 1998, Fleming et al. 1998). Over 121 March flies (Family Tabanidae) and blowflies were found in the stomach of a trapped dingo (even though flies are not regarded as food) and were observed to pester trapped dingoes (Newsome et al. 1983).
Oedema is a common indication of potential ischemia and is frequently observed after trapping in padded leg-hold traps (Andelt et al. 1999), yet in some cases animals recover after release in a few days with no further indications of injury (Saunders et al. 1984). Oedema of varying degrees is seen in foxes captured with treadle-snares (Figure 6a) and Victor Soft-Catch traps (Figure 6b).
During the obstruction of blood flow, cellular production of ATP may use glycogen metabolism and creatine phosphate until depleted. Glycogen is broken down to yield pyruvate and lactate and causes a decline in pH, which will then limit phosphofructokinase activity and further reduce the potential to develop ATP stores (Harris et al. 1986). There is a good relationship between the depletion of skeletal muscle ATP stores and the extent of ultimate muscle necrosis (Walker 1991). Using laboratory rodents subjected to periods of ischemia, reflow of blood into capillaries was inhibited after two hours and upon release of the tourniquet declined for a further 90 minutes (Forbes et al. 1995). Restricted blood flow to limbs will not return to levels seen before ischemia and tissue damage continues for a period thereafter. The mechanism responsible may relate to the obstruction of capillaries with leukocytes (Schmid-Schönbein 1987), neutrophil mediated injury in tissue (Schraufstatter et al. 1984, Ellard et al. 2001), the swelling of endothelial cells (Harris et al. 1993) or free-radical mediated damage (Walker 1991).
The sudden return of circulation initiates the conversion of injured tissue that may have shown superficial oedema to necrotic tissue over a few days and may take a protracted period to develop completely (Walker 1991). Tissue pressure of 50 mm Hg represented a critical threshold for human peripheral nerves at which there will be acute damage (Gelberman et al. 1983). Pressures of 300 mm Hg cause almost total occlusion of blood flow in the limbs of monkeys (Klennerman et al. 1977). The application of tourniquet cuffs was shown to damage the sciatic nerve of dogs and although the degree of impairment differed between individuals, full recovery was shown to take up to 6 months (Rorabeck et al. 1980). Using lower pressures of 200 mm Hg for two hours, temporary peripheral nerve conduction and blood flow was occluded and a degree of nerve injury was most pronounced a week after treatment and diminished in severity over a six week period (Nitz et al. 1986). An important implication for welfare outcomes from trapping is that after a period of ischemia, gross pathology will be visible only well after blood flow is restored (Walker 1991). The incidence of debilitation cannot be known unless the fate of animal is followed subsequent to release or detailed veterinary investigations are made of affected tissues prior to release.
Figure 6. Typical oedematous swelling in the paws of foxes restrained by a treadle snare (on the left front leg) (a) and the Victor Soft-Catch #3 trap (on the right front leg) (b) for unknown durations.
Secondary, chronic and pathological pain
When an animal is trapped there may be pain associated with the initial closure of the trap and ongoing pain from the trap mechanism (eg. clamping pressure). Pain may arise from injury due to struggling and continue subsequent to an animal’s release, as it does not necessarily abate upon the removal of a painful stimulus. Pain is probably best considered to be both a stressor and a form of a variable pathology that may have a complex feedback mechanism. Chronic pain may be variously defined as pain that is perceived subsequent to healing or pain that has no useful purpose (discussed by Rutherford 2002). Chronic pain is not an extension of acute nociceptive pain, but may be an evolving process where injuries produce a chain of pathogenic mechanisms that initiates another (Covington 2000). Pain may not arise immediately, but may follow some time after trauma and exertion (Marchettini 1993). For example, muscle pain may arise from a state of nociceptor sensitisation and be associated with strenuous muscle activity (eg. post-exercise muscle pain) and associated with myopathic weakness and an increase in serum muscle enzymes. Pain that arises after neurological damage (neuropathic pain) after trauma and the development of inflammation (Carstens et al. 2000) can be chronic and associated with abnormal nociception and amplification of pain.
After nerve damage associated with trap injury, the cessation of physical trauma cannot be assumed to eliminate pain as acute pain may sensitise and/or facilitate the development of chronic pain mechanisms (Covington 2000). Pain can become more intense due to restricted venous return as the wound area becomes engorged with blood and the veins are occluded. The pressure from this swelling may directly activate pain receptors and the stimulus producing the pain cannot be removed by restoring blood flow (Gregory 2005). Animal models of neuropathic pain have been developed in rats by the placement of loose ligatures around the sciatic nerve or dorsal roots. When the limb becomes oedematous it is often held in the air and animals develop long lasting and extreme sensitivity to heat and mechanical stimulus beyond the area of nerve damage (Bennett et al. 1988, Kim et al. 1992).
Kuehn et al. (1986) recorded that up to 3% of grey wolves chewed at their trapped limbs irrespective of whether traps were toothed or offset. Self-mutilation is frequent in raccoons (Procyon lotor) captured in padded and unpadded leg-hold traps (Berchielli et al. 1980). Dingoes were observed to gnaw at their trapped leg, sometimes biting off extremities (Newsome et al. 1983) and this is common in coyotes (Balser 1965). In other studies, injury sustained by the trapped limb was possibly produced as the animal gnawed at the device, implying that it may not always be self-directed (Stevens et al. 1987). A fox cub was found to mutilate its digits below the point at which it was held by a Victor #3 Soft-Catch trap (C.A. Marks, unpublished data). Self-mutilation of trapped feet was observed in 2/10 coyotes trapped in modified Victor Soft-Catch traps (Houben et al. 1993). Using off-set and laminated Bridger #3 traps to capture coyotes, 2/27 were also found to have chewed their foot pads (Hubert et al. 1997). Self mutilation was observed in 2/107 pumas captured in leg-hold snares after lower leg bones had been broken (Logan et al. 1999). Raptors were found to self-mutilate following traumatic nerve injury (Holland et al. 1997), yet their propensity to do this in traps is unknown.
The relationship between nerve damage and self-mutilation is still unclear and previously some authors proposed that it occurs as a way for animals to shed impaired and insensitive appendages (autotomy) (Rodin et al. 1984), although this explanation has not found wide support. It has also been suggested that the reasons for self-mutilation of animals trapped in leg-hold traps may be because of progressive limb desensitisation (Gregory 2005) and may imply that injury is not self-directed. However it appears likely that pain and nerve damage is most likely the primary stimulus that directs self-mutilation. When the sciatic nerves of rats were severed, 80% were observed to self-mutilate the desensitised area (Blumenkopf et al. 1991) and this was also observed in 91% of subjects in another study (Wall et al. 1979) There is evidence that genetically determined variation in rates of autotomy/self-mutilation occurs within some species (Coderre et al. 1986). Self-mutilation has become an important marker of pain in the assessment of analgesics (Willenbring et al. 1994). In mice, self-mutilation begins at the toes, after which biting progresses further up the limb. Anaesthetic applied to the limb before nerve damage prevents the onset of self-mutilation and this implies that pain perception is important in the initiation of this behaviour. In chronic post-traumatic situations in humans, most commonly following traumatic brachial plexus avulsion, patients have attempted to persuade others to amputate the limb in the hope of relieving unremitting neuropathic pain. In these cases the prime motivation is almost always relief of pain rather than merely the removal of dysfunctional anatomy (Bonney 1959).
6.2.2 Anxiety and fear
Fear is an emotional response to a potentially harmful stimulus and is sometimes separated from anxiety which is defined as an emotional response to a stimulus that predicts a potentially harmful or unpredictable environment (Casey 2002). In this definition, fear is elicited by an explicit, threatening stimulus and subsides shortly after offset of that stimulus (Davis et al. 1997). Anxiety may be a more generalised and may last for a long period once activated (Davis et al. 1997). It is a different state to that of fear as it is mostly related to anticipation or dread in the absence of external triggers (Gregory 2005). Gregory (2005) lists a range of situations that produce fear, including capture, exposure to unfamiliar objects and odours, sudden movement, separation from companions, aggressive encounters, exposure to predators and predator related cues. This emphasises that fear and anxiety are probably experienced in response to a broad range of stressors encountered during trapping. Sudden and violent alarm (eg. startle response), apprehension and frustration may be states related to fear and anxiety and are deemed to be psychological stressors (Jordan 2005) (Figure 7). They are motivators induced by the perception of danger and each has survival value if life expectancy of animals can be increased if danger is avoided (Boissy 1995). In monitoring the environment for threats, an animal will respond with fear if there is a large discrepancy between observed and expected stimuli (Archer 1979). Fear in animals is believed to give way to either defensive or avoidance behaviours as a way to protect them from potentially harmful situations (McFarland 1981).
Fear and anxiety can become pathologic when the stressors are intense or prolonged. Tissues can be damaged by short-term immobilisation and even emotional or social stress. Immobilisation or restraint in the absence of other stressors can induce myocardial lesions and affect tissue integrity in vital organs (Sanchez et al. 2002). This finding challenges previous beliefs that stress operates over longer periods to cause pathology; short-term restraint may have greater implications for the welfare of animals than previously thought. When in contact with humans, Arctic foxes express fear that is well known to be associated with an increase in stress hormones (Kenttämines et al. 2002). Domesticated animals have a reduced functional activity of the pituitary-adrenal system, a decreased total glucocorticoid level in blood and, from in vitro studies, appear to produce less adrenal hormones and basal levels of ACTH (Trut et al. 2004), yet silver foxes that have been bred to be resistant to stress (Belyaev 1978; 1979) will display rapid stress-induced hyperthermia (SIH) when in close proximity to humans (Moe and Bakken 1997; 1998, Trut et al. 2004). This has also been observed in laboratory rodents. In each species, it has been related to the induction of the HPA and sympathetic adrenal-medullary system (Moe et al. 1997); there is little increase in physical activity associated with SIH in foxes (Moe and Bakken 1997 and rodents (Kluger et al. 1987). All vertebrate species probably possess specific receptor sites for benzodiazepine drugs, which influence states of anxiety (Rowan 1988), and diazepam has been used successfully to manage SIH in laboratory rodents and foxes (Moe et al. 1998), reinforcing that anxiety or fearful states initiated solely by the presence of humans are probably responsible for SIH.
Figure 7. Anxiety, fear or frustration? A dingo captured with a Victor Soft-Catch #3 trap instrumented with an activity monitoring data logger (metal box fixed to chain) after approximately 10 hours of confinement (after Marks et al. 2004).
6.2.3 Capture myopathy and exhaustion
Capture myopathy is an acute degeneration of muscle tissue that may arise from capture and restraint, especially when associated with intense muscular exertion (Hulland 1993). It is a condition variously named as transport stress, capture stress, degenerative myopathy, white muscle disease or exertional rhabdomyolysis. It is primarily a disease of wild and domestic animals and is most commonly reported in birds and mammalian taxa such as macropods, deer, cetaceans, seals, rodents and primates (Williams et al. 1996). Trauma or compression of muscles due to physical injury, long-term confinement in the same position, strenuous activity and constriction of blood flow or hyperthermia are among a number of stressors that can lead to muscle damage (Vanholder et al. 2000). The disease is initiated when exertion during anaerobic glycolysis produces low muscle pH associated with the accumulation of lactic acid in muscle cells. Cellular enzymes such as CK, AST, and LDH are released into the blood stream along with free radicals that can overwhelm the protective and corrective antioxidant defence mechanisms (Viña et al. 2000). Diagnosis of exertional myopathy is usually based upon history of susceptibility, clinical signs and elevation in AST, CK and LDH (Dabbert et al. 1993). Upon the death and disintegration of muscle tissues, myoglobins (that resemble haemoglobin) are released and can damage the kidney and the lungs (Wallace et al. 1987, Vanholder et al. 2000) and when severe, urine may be discoloured dark brown. Acute renal failure may result from a combination of acidosis and ischemia in concert with myoglobin deposition in the glomeruli (Wallace et al. 1987). Normally, free myoglobin is bound to plasma globulins, but massive release of myoglobins will exceed the capacity of plasma proteins to bind them. Short and intensive bursts of activity may contribute more to the onset of capture myopathy than prolonged but less intense activity (Beringer et al. 1996).
There are four general appearances of the disease that have been best described in livestock:
- Hyperacute (peracute) capture myopathy – associated with very sudden onset and death due to shock and vascular collapse;
- Acute capture myopathy – where animals survive for hours or days;
- Sub-acute capture myopathy - Ruptured muscle syndrome occurs within days to weeks and the animals develop painful movement due to muscle rupture;
- Chronic – associated with death that supervenes after a second capture attempt due to predisposition to cardiac arrest and arrhythmias due to capture myopathy or pathogen and parasite related diseases (Spraker 1993, Rendle 2006).
While not normally a disease commonly associated with carnivores and dogs in general (Aktas et al. 1993), capture myopathy was described in a red fox (Little et al. 1998) and in river otters (Lutra canadensis), where clinical signs included depression, anorexia and shock, although it was not the sole cause of death (Hartup et al. 1999). Capture myopathy has been reported for 11 species of macropods in Australia with either debilitation or death being the outcome (Shepherd et al. 1988). No evidence of cardiac necrosis or renal damage was found in a study on red kangaroos (Macropus rufus) although skeletal muscle necrosis and myoglobinuria was found in many (Shepherd 1983, in Shepherd 1988). An attempt to capture macropods in Australia has been documented to result in 37% (Keep et al. 1971) and 100% (Shepherd et al. 1988) mortality due to capture myopathy. This has lead to the development of trapping and immobilisation techniques for small macropods that are specifically designed to avoid injury and capture myopathy (Coulson 1996, Lentle et al. 1997). It has been suggested that long-legged birds are more susceptible (Hanley et al. 2005) and appears to be the case with emus (Tully et al. 1996). The presentation and clinical signs of the disease appear to vary and may be species-specific. Three roe deer were captured in drive nets, restrained and placed in transport boxes and then translocated to an enclosure where they were observed to die 48 hours, 72 hours and 8 days after capture, possibly due to a second stress episode (Montane et al. 2002). When using ‘drop nets’ it was estimated that between 6-16% of white-tailed deer (Odocoileus virginianus) suffered capture myopathy. Sedating and blindfolding animals and limiting the noise associated with handling was shown to reduce capture myopathy by 50% (Connor et al. 1987) and probably demonstrates the importance of handling, light and acoustic stressors in managing this disease. The use of traps that reduce handling and processing times and overall exertion were found to significantly decrease the incidence of capture myopathy compared to the use of net guns (Beringer et al. 1996).
Animals suffering from capture myopathy may be debilitated by scarring of skeletal or cardiac muscles, which may cause them to appear slower or less alert after release. This may make animals more susceptible to predation or to other stressors that can cause their death weeks or months after capture (Hulland 1993). The prognosis is poor for animals that have clinical signs of capture myopathy, especially if released immediately (Rogers et al. 2004, Hanley et al. 2005). In whooping cranes (Grus americana), routine capture and handling caused exertional myopathy and treatment was unsuccessful (Hanley et al. 2004). Some success has been reported in a range of shorebirds that were rendered unable to stand, walk or fly, yet this took up to 14 days of intensive supportive care (Rogers et al. 2004). Similarly, muscle tissue killed by myopathy in quokkas (Setonyx brachyurus) was found to regenerate after 5-8 weeks (Kakulas 1966). Selenium (0.06 mg kg-1 as sodium selenite) and vitamin E
(0.45 mg kg-1 as d-α tocopherol acetate) was shown to be beneficial in protecting and assisting recovery of myopathy conditions in livestock (Viña et al. 2000). Treatment of northern bobwhites (Colinus virginianus) significantly increased the survival of birds compared to a placebo and this was attributed to a reduction in pathology associated with capture myopathy (Abbott et al. 2005). Given that many wildlife species will be intractable to long-term captivity, the practicality of providing supportive care in the field to non-target species suspected of suffering myopathy is questionable.
Exertion until exhaustion will have species-specific consequences; dogs appear to be able to endure exercise stress until exhaustion without severe metabolic acidosis and were found to have a resting cardiac output 30% greater than pigs, which increases to 60% greater during steady-state exercise (Hastings et al. 1982). Wombats caught in steel-jawed traps that are tethered to solid objects will often respond by continuous digging with their unrestrained limb until physical exhaustion characterised by lethargy and unresponsiveness is seen (C.A. Marks, personal observations), which is clearly associated with poor welfare (Anon 2007). Unfortunately, there are few data available concerning the susceptibility of many wildlife species to capture myopathy and their fate subsequent to release.
6.2.4 Hyperthermia and hypothermia
Heatstroke (hyperthermia) occurs when the mechanisms responsible for heat loss are overwhelmed, particularly in the absence of freely available water in species such as domestic dogs. In dogs the disease is characterised by marked elevation in core body temperature resulting in widespread hepatic and gastrointestinal cellular damage as body temperature approaches 42oC with vascular collapse, shock and death (Bosak 2004). At body temperatures in excess of 41oC, domestic dogs are unable to maintain thermal equilibrium and collapse, and neurological symptoms are evident above 42.5oC (Andersson 1972). After 30-60 min of moderate exercise on a treadmill (4 km h-1 at an 8% gradient) at air temperatures between 3742oC, Alsatian dogs became distressed and attempted to escape (Bedrak 1965).
The early stages of heatstroke in dogs are characterised by hyperthermia, tachycardia, depression, vomiting, diarrhoea and dehydration (Krum et al. 1977). Independent of thermal stressors, anxiety and stress can induce hyperthermia in silver foxes within 5 minutes (Moe et al. 1997). Heat stress was associated with the death of animals in traps despite the use of the TTD containing diazepam (Balser 1965). Elevated body temperature was associated with capture deaths in black bears caught with foot-snares (Balser 1965).
Most small vertebrate species, including arid adapted mammals and reptiles, will become thermally stressed when ambient temperatures exceed 40-45oC in traps and prolonged exposure may result in death (Hobbs et al. 1999). Some bandicoots that are found in mesic environments such as the eastern barred-bandicoot (Perameles gunnii) are similarly unable to tolerate ambient temperatures > 35oC (Larcombe et al. 2006). Common wombats reduce heat loss during winter by active periods of feeding followed by refuge in a burrow where their heat loss is reduced. They have difficulty in maintaining a constant body temperature when ambient temperatures exceed 25oC and show severe thermal stress when exposed to temperatures above 30oC (Brown 1984). Arid adapted wombats such as the southern hairy-nosed wombat avoid high temperatures in summer by selecting cooler parts of the evening to forage, and they appear to be poor at regulating their body temperature (Wells 1978). The echidna is unable to manage ambient temperatures > 35oC and relies upon shelter in burrows to maintain a body temperature below a fatal body temperature of 38oC (Brice et al. 2002).
Hypothermia is a condition where the animal’s body temperature drops below that required for normal metabolism. Signs of hypothermia include shivering, lethargy, muscle weakness, stupor, coma and death if severe (Kayser 1957). Rapid chilling is associated with pain and discomfort, especially to the extremities, and reperfusion pain when full circulation is restored (Gregory 2005). Some trapped species have been recorded to die as a result of hypothermia in North America (Mowat et al. 1994) and it has been listed as a possible cause of death for trapped Australian species (Fleming et al. 1998). Overnight temperatures < - 8 oC were found to be associated with risk of freezing injury in lynx (Mowat et al. 1994). In sub-zero temperatures the common wombat appears to be dependent on access to burrows in order to avoid hypothermia (Brown 1984), although southern hairy-nosed wombats may be mildly tolerant to hypothermia (Wells 1978). Poorly insulated shelters can cause death through hypothermia in common wombats housed in zoos (Marks 1998b).
6.2.5 Impact on dependent young and reproduction
The welfare, growth and reproductive performance of agricultural, laboratory and zoo animals can be negatively affected by fear and anxiety (Boissy 1995). Trapping stress, injury and death may cause: 1. ejection of pouch young; 2. abortion; 3. the death of dependent offspring and; 4. welfare impacts that arise from prenatal stress altering HPA responsiveness in utero and consequent effects upon the behaviour of offspring. The period of reproductive activity corresponding to gestation, birthing period and maintenance of target and non-target young gives some indication of the periods that correspond to possible impacts of trapping upon reproduction and offspring (Figure 8).
The trapping of animals with offspring that are dependent upon lactation, food and maternal or paternal care is a possible outcome when traps are used during times corresponding to breeding, birth and care of target and non-target young (Sharp et al. 2005a; 2005b). In dogs and foxes, lactation is vital to the survival of cubs maintained within the natal den before they begin to accept prey (Tembrock 1957). The care of dependent young is also highly dependent upon a wide range of roles fulfilled by the adults of different species, such as egg incubation, provision of shelter, protection from predation, provision of body heat and potentially the maintenance and protection of young past early dependence (Clutton-Brock 1991).
Dingoes appear to breed only once each year in the wild, yet births in the eastern highlands of Victoria were estimated to occur over a seven month period from March to September with a breeding peak from June to August (Jones et al. 1988). Male dingoes were found to have either a low intensity testicular cycle (Jones et al. 1988) or none at all (Catling 1979). Breeding in domestic dogs is variable in timing and can occur more than once each year (Christie et al. 1971), with males being fertile throughout the year (Kirk 1970, in Jones and Stevens 1988). The red fox will produce a single litter each year after a 52-53 day gestation period (Lloyd et al. 1973, Ryan 1976, Coman 1983). In Australia, pregnancies in the fox have been reported to range from June to October in foxes taken from a range of habitats across New South Wales (Ryan 1976) and from July to October in Canberra (35oS) (McIntosh 1963). In a study in western New South Wales (32-33oS), the timing of mating and births varied from 7 weeks in 1995 to 3-3.5 weeks in 1996, and the earliest evidence of oestrus was detected on the 14th June (McIIroy et al. 2001).
The bandicoot genera, Isoodon and Perameles, contain highly fecund species, with multiple births each year with short inter-litter intervals. The macropods (Macropus and Wallabia) breed year round (Menkhorst 1995). Along with brushtail possums, they have a much larger inter-litter period (> 200 days) and while brushtail possums have a major autumn and minor spring breeding season, breeding may occur year round (How 1988). Wombats have been shown to breed throughout the year, although in Victoria there appears to be a cluster of births in summer (Skerratt et al. 2004). Lyrebirds and emus breed from May through to October and ravens between July and September. The period of care provided by the male emu for chicks has been recorded to last as long as 18 months and may be a period of three to four months for ravens (Schodde et al. 1990). Although seasonal breeders can have more predictable reproductive cycles, the period of maternal care necessary to ensure the survival of juvenile offspring is difficult to define with any precision.
Ejection of young and abortion
In macropods, the ejection of pouch young due to stress or predator avoidance is a unique strategy to assist in the survival of the mother when stressed (Coulson 1996). The ejection of pouch young in response to stress has been observed in eastern-grey kangaroos (in Coulson 1996) and swamp wallabies (Robertshaw et al. 1985). Stress-induced inhibition of prolactin secretion, resulting in diminished progesterone concentrations, might be the chief cause of reproductive failure and abortion in red foxes (Hartley et al. 1994). Stress induced abortions have been noted as a consequence of trapping stress, yet may not occur immediately. For instance, a puma injured during trapping with a leg-hold snare aborted 3-4 days after capture but was only recorded because it was closely monitored (Logan et al. 1999).
Prenatal stress in the last third of pregnancy induced by brief handling affected adrenal weight and adrenocortical function in blue fox offspring (A. lagopus) (Braastad 1998). This follows a general observation that anxiety during pregnancy can affect the corticosterone response to stress (Vallee et al. 1997), although the welfare implications of this finding are not clear.
Figure 8. Period of gestation following mating and potential birth season and period of care for dependent young (lactation and maternal care) for target and non-target species (Strahan 1984, Lee et al. 1985, Tyndale-Biscoe et al. 1987, Hayssen et al. 1993, Menkhorst 1995, Temple-Smith et al. 2001, Menkhorst et al. 2004).
6.2.6 Dehydration and starvation
If obligatory loss of water (eg. panting, salivation, urination etc) is not replaced by water ingestion, raised extracellular fluid osmolarity and reduced extracellular fluid volume will rapidly cause a state of cellular and extracellular dehydration (Ramsay et al. 1991) and may contribute significantly towards the onset of hyperthermia in hot environments. In dogs, 24 hour water deprivation results in a steady rise in plasma osmolarity and an increase in plasma vasopressin without a decline in urine volume because water excretion is required to eliminate sodium (Thrasher et al. 1984). Black bears captured in Aldridge snares had blood biochemistry profiles attributed to greater exertion, muscle damage and dehydration compared to individuals captured by remote activated tranquilising collars (Powell 2005). Grizzly bears had higher N:L ratios, as well as increased concentrations of Na and Cl that were attributed to dehydration due to water deprivation during 2-23 hours of captivity, which was probably aggravated by intense activity (Cattet et al. 2003). Increased CK, PCV, ALB,