Considerations for furbearer trapping regulations to prevent grizzly bear toe amputation and injury

Science and adaptive management form crucial components of the North American model of wildlife management. Under this model, wildlife managers are encouraged to update management approaches when new information arises whose implementation could improve the stewardship and viability of wildlife populations and the well-being of animals. Here we detail a troubling observation of several grizzly bears with amputated toes in southeast British Columbia and assemble evidence to inform management strategies to remedy the issue. During the capture of 59 grizzly bears, we noticed that four individuals (~7%) had amputated toes on one of their front feet. The wounds were all healed and linear in nature. Further opportunistic record collection revealed that similar examples of amputated toes occurred beyond our study area, and that furbearer traps were frequently responsible for toe loss. We found evidence that seasonal overlap between the active season for grizzly bears and the fall trapping seasons for small furbearers with body grip traps and for wolves with foothold traps may explain the issue. Multiple options to reduce or eliminate this issue exist, but have varying degrees of expected efficacy and require differing levels of monitoring. The most certain approach to greatly reduce the issue is to delay the start of the marten trapping season until December 1, when most bears have denned, instead of opening the season on or prior to November 1, when more than 50% of bears are still active. Additionally, innovative solutions, such as narrowing trap entrances to exclude bear feet while still allowing entrance of target furbearers, have the potential to minimize accidental capture of bears, but the effectiveness of these approaches is unknown. Experimental evidence suggested that better anchoring traps was not a viable solution. Solutions that do not involve season changes will require monitoring of efficacy and compliance to ensure success.

ABSTRACT Science and adaptive management form crucial components of the North 26 American model of wildlife management. Under this model, wildlife managers are encouraged to 27 update management approaches when new information arises whose implementation could 28 improve the stewardship and viability of wildlife populations and the well-being of animals. 29 Here we detail a troubling observation of several grizzly bears with amputated toes in southeast 30 British Columbia and assemble evidence to inform management strategies to remedy the issue. 31 During the capture of 59 grizzly bears, we noticed that four individuals (~7%) had amputated 32 toes on one of their front feet. The wounds were all healed and linear in nature. Further 33 opportunistic record collection revealed that similar examples of amputated toes occurred 34 beyond our study area, and that furbearer traps were frequently responsible for toe loss. We 35 found evidence that seasonal overlap between the active season for grizzly bears and the fall 36 trapping seasons for small furbearers with body grip traps and for wolves with foothold traps 37 may explain the issue. Multiple options to reduce or eliminate this issue exist, but have varying 38 degrees of expected efficacy and require differing levels of monitoring. The most certain 39 approach to greatly reduce the issue is to delay the start of the marten trapping season until 40 December 1, when most bears have denned, instead of opening the season on or prior to 41 November 1, when more than 50% of bears are still active. Additionally, innovative solutions, 42 such as narrowing trap entrances to exclude bear feet while still allowing entrance of target 43 furbearers, have the potential to minimize accidental capture of bears, but the effectiveness of 44 these approaches is unknown. Experimental evidence suggested that better anchoring traps was 45 not a viable solution. Solutions that do not involve season changes will require monitoring of 46 efficacy and compliance to ensure success. 47

IDENTIFYING AN ISSUE 51
Grizzly bears (Ursus arctos) are wide-ranging mammals that are highly motivated to 52 rapidly ingest high-energy foods in preparation for denning (McLellan 2011). In the Rocky 53 Mountains of southern Canada, grizzly bears occupy large home ranges (200-1300 km 2 ) where 54 they search for mates and seasonal foods (Graham and Stenhouse 2014). Key natural foods for 55 grizzly bears include high-calorie fruits, abundant herbaceous vegetation, ungulates (killed or 56 scavenged), ants, and roots (McLellan and Hovey 1995, Munro et al. 2006). In addition to these 57 natural foods, grizzly bears are well known to take advantage of poorly managed human foods 58 such as residential fruit trees, garbage, roadkill, grain, and livestock (Craighead and Craighead 59 1972, Lamb et al. 2017, 2019, Morehouse et al. 2020). This food-motivated behaviour is 60 generally adaptive and allows grizzly bears to increase their weight by about 35% in the 6-8 61 months of non-denning season (McLellan 2011), but bears can also be attracted to, and allowed 62 to feed on, human-sourced foods which can create human and bear safety issues (Lamb et al. 63 2020). 64 The challenges to human-bear coexistence are exacerbated when high-density grizzly 65 bear populations co-occur in human settled areas with abundant natural foods and accessible 66 human-sourced foods. One area that fits this description well is the southeast corner of British 67 Columbia (BC), Canada. This region is home to high-density grizzly bear populations, 68 productive habitat, and growing rural and urban communities. In response to recent grizzly bear 69 population declines, ecological trap dynamics, and high human-caused mortality rates in this area 70 (Lamb et al. 2017(Lamb et al. , 2019, 2020), we initiated a radio collaring project to better understand sources 71 of grizzly bear mortality and examine how animals' use of the landscape influenced population 72 dynamics. Between 2016 and 2020, we captured and radio collared 59 grizzly bears in the Elk 73 Valley (near Fernie, BC, see Fig 1A). Captures  (~7%) were missing some of the toes on one of their front feet (Fig 2). In all cases, the toes lost 84 were adjacent and terminated in a linear plane, and the injury had healed. At the time of their 85 capture, these animals were all between the ages of 4 and 12 years old. Three of the four 86 captured bears with amputated toes were male, while the larger collar sample had slightly more 87 females (n=31) than males (n=28). The female with amputated toes was captured with a yearling 88 cub, suggesting that amputated toes did not preclude reproduction. Based on radio collar 89 telemetry data, these animals moved in a similar fashion to bears with all their digits. While toe 90 loss did not appear to impact their locomotion, it may have influenced conflict behaviour, which 91 was common in these animals; three of the four animals with amputated toes were involved in 92 human-bear conflicts. One of these four bears was killed shortly after capture while breaking into 93 a rancher's calf pen, another was likely involved in an attack on a human, and a third was 94 captured by conservation officers in late fall after complaints of bear issues on a farm. While the 95 sample size for this conflict behaviour is small, less than one third of the bears with all their toes 96 (n=55) died or were involved in conflict during the time they were collared. The reason for 97 elevated conflict behaviour observed in the bears with amputated toes could be related to 98 challenges feeding on natural foods at certain times of year, such as digging roots during spring 99 and fall (McLellan and Hovey 1995), or these bears could be naturally bolder individuals and 100 thus more likely to become both trapped and involved in future conflict. Alternatively, the 101 pattern of increased conflict could be an artefact of the small sample sizes we had available 102 (n=4). Nevertheless, while we had identified a concerning trend of amputated bear toes but the 103 healed wounds left us little evidence for the cause of the toe loss. 104

SEARCHING FOR CAUSE 106
We first consulted with veterinary professionals to determine if the amputated toes could 107 be congenital (i.e., lost before birth and due to natural causes). After review of multiple photos 108 and X-rays (Fig 2) there was little evidence to suggest that this was congenital. On the contrary, 109 strong evidence for injury post-birth included clear evidence of toe bone fracture and healing, 110 and consistent linear wounds suggesting amputation related to human sources. As a result, we 111 proceeded to search for a human-caused source of the issue. 112 Southern British Columbia is fortunate to have a long history of people interacting with 113 grizzly bears through research, human-wildlife conflict work, hunting, guide outfitting, trapping, 114 and simply living in bear country. After identifying the amputated toe issue, we spoke with First 115 Nations, scientists, conservation officers, wildlife managers, guide outfitters, and trappers, all 116 with lifetimes of experience on this land. A common topic in these discussions was grizzly bears 117 being incidentally caught in baited traps set for furbearers, particularly in traps set for marten 118 (Martes americana) and wolves (Canis lupus). Indeed, non-target captures stemming from legal 119 trapping activities have been reported for many species. Cougars (Puma concolor) have been 120 incidentally captured in foothold traps or snares set for wolves and bobcat (Lynx rufus) ( to capture marten and weasels, and they are typically set at the mouth of a baited wooden box 133 that is affixed to a tree at a height of approximately 1.5 meters from the ground. 134 In the spring of 2018, we deployed four body grip traps, complete with their boxes and 135 bait (usually a small piece of meat), on trees to see if bears would investigate these traps. We 136 wired the traps in such a way that they could fire but not close and trap the bears. We monitored 137 the traps with remote cameras for approximately two weeks to see if bears would visit and spring 138 the traps. Grizzly bears visited all four traps and sprung two of them. Pictures and videos showed 139 bears investigating the traps and manipulating the boxes with their paws (Fig 3). Video showed 140 that one of the traps was set off with a bear's nose, but the cause of the other trap being set off 141 could not be determined because we only had pictures, not video, of bears investigating the trap. 142 Although this was a small trial with low sample sizes, younger bears seemed to investigate and 143 set off traps more often than older bears. Even with the small sample, it was clear that baited 144 traps attracted bears and that bears set off the traps as they tried to get the bait. We heard from 145 some trappers that they voluntarily delayed the start of their marten trapping season (which 146 legally opens November 1, Province of British Columbia 2021) to avoid having bears wreck 147 their trap sets, further confirming that bears are attracted to these common sets. 148 We became more convinced that furbearer traps were the cause of the amputated toes as 149 we accumulated anecdotes and photo evidence of grizzly bear feet stuck in traps (Fig 2). Photos 150 came from hunters, remote cameras on ranches, and conservation officers, and they all suggested 151 the same mechanism. A body grip trap was affixed just behind the toes and amputation was 152 underway. Beyond identifying a likely cause for the amputated toes, the photo evidence and 153 information from people out on the land revealed that bears with traps stuck on their feet and 154 amputated toes was an issue beyond the extent of our study area. In addition to the records of 155 body grip traps on bears' feet in southeast BC (also known as the "Kootenays"), photos of a 156 We were also aware of multiple reports of grizzly bears being caught in foothold traps set 160 for wolves, and we believe this is another possible source of toe loss. Between 2010 and 2020, at 161 least five grizzly bears in southern BC were caught in wolf foothold traps (with the trap often 162 closing right behind the toes) and had to be released by conservation officers and scientists. 163 However, in the records we were able to collect, body grip traps were the most common trap on 164 bears' feet where it was clear the trap was the cause of toe loss. In one case, it is suspected that a 165 bear's foot became stuck in a body grip trap set on the ground for skunks on private property. in the Selkirk and Purcell Mountains, but no evidence of toe loss due to incidental grizzly bear 179 capture in footholds was reported. Likely because bears were either released from the traps or 180 killed. None of the bears captured in the Flathead Valley had amputated toes; however, a hunter 181 killed a bear just north of the Flathead study area with a body grip trap on its foot (Fig 2E). 182 Results from these adjacent studies did not reveal any general insights into the observed spatial 183 distribution of bears with amputated toes. This is not surprising given that traplines in British 184 Columbia are distinct areas trapped exclusively by registered trappers, and variation in trapping 185 intensity, target species, and timing creates variable exposure of bears to body grip traps across 186 the landscape. Additionally, unsecured attractants such as apple trees, garbage, roadkill, and 187 livestock may increase bear use of low elevation areas and delay den entry, perhaps increasing 188 the chance of overlap with trapping activities in these areas. 189 Observations of bears stuck in body grip traps and bears with amputated toes have mainly 190 been documented in the past 15 years and are not restricted to a single area or trapline, but rather 191 occur broadly at low frequency across the Kootenays and beyond. This increased detection of 192 bear toe loss could be related to an increase in grizzly bear research and monitoring; however, 193 opportunities to identify these issues have been abundant for several decades as many bears were 194 captured by conservation officers, harvested by hunters, and monitored for research purposes in 195 the Flathead. It is possible that in recent years the increased use of stronger traps, as mandated by 196 evolving humane trapping standards, has created a situation where bears now have more 197 difficulty extracting themselves from body grip traps when captured. 198 We summarized the den entry and exit dates for 61 animal-years in the Elk Valley to 199 assess when bears were denning in the fall. We found the median den entry date (i.e. the date 200 when 50% of bears had denned) was November 6, and the 95 th quantile of den entry was 201 November 22, revealing a critical 3-week period of overlap between active bears and the 202 trapping season. We found no overlap between den emergence and the end of the trapping 203 season; thus, we focus on fall trapping as the source of the issue. More support for the idea that 204 the concurrence of the fall active season for bears and the beginning of the trapping season is the 205 main issue is that we have not seen similar problems of amputated toes in black bears, who den 206 several weeks earlier than grizzly bears and would thus be less exposed to traps. 207

HONING A SOLUTION 209
Most trapping of furbearers is done in winter when fur is prime (i.e., underfur is dense 210 and guard hairs are long) and most valuable. Given that grizzly bears hibernate for the winter, 211 trappers should generally be able to avoid accidently catching bears. However traps set for marten during the month of November must be enclosed in a box with an opening no 234 larger than 3 ½ inches. This constricted entrance was thought to be narrower than most bear 235 paws. The license condition also recommended that, prior to December 1, trappers should use 236 these same boxes to enclose similarly sized killing traps set for other species on dry ground. The 237 efficacy of the modified enclosure entrance in eliminating bear toe loss has not yet been tested, 238 but these modified marten trap boxes are believed to be sufficient to ensure grizzly bears are not 239 able to access a set trap. Because there is more uncertainty in the effectiveness of the modified 240 enclosure approach than there is in delaying the start of the trapping season, we recommend that 241 compliance and efficacy is monitored to ensure this intervention is effective and uptake is high. 242 Such monitoring would allow for adaptive management and changes to the approach as needed 243 to ensure a successful outcome for bears and trappers alike. 244 Additional approaches have been suggested to help resolve the issue, but we do not 245 recommend their implementation. For example, one option is to sufficiently anchor traps so a 246 bear that is incidentally caught can hopefully pull its foot free. To address the viability of this 247 solution we conducted a small experiment using front appendages of deceased bears and four 248 types of traps approved for marten trapping in British Columbia (Province of British Columbia 249 2021). Our experiment consisted of three phases: the first used before and after X-rays to assess 250 whether traps could immediately break bear toe bones, the second assessed whether bears could 251 easily pull their feet free from traps (which we simulated by solidly anchoring the trap and 252 having an 86 kg person attempt to free the feet by pulling on the leg with maximum human pull), 253 and the third assessed how much pull (kg) was required to free the feet. We gathered the front 254 appendages from one adult and two cub grizzly bears killed in the Elk Valley during fall 2021 255 and tested four approved body grip traps: 1) Bélisle Super X 120, 2) Northwoods 155, 3) 256 Sauvageau 2001-5, and 4) Sauvageau 2001-6. 257 Results from our experiment suggested that traps do not immediately break bones, but 258 bears were unable to easily extract themselves from these traps. X-ray results confirmed that 259 setting traps off on adult or cub feet did not break or fracture any bones. Despite using maximum 260 human effort and both sustained and jerking pulls, we were only able to free adult grizzly feet 261 from traps 20% (95% CI: 7-33%) of the time, and cub feet 63% (95% CI: 53-72%) of the time. 262 For 80% of our attempts with adult feet and 37% of our attempts with cub feet, we were unable 263 to free the feet from the traps despite repeatedly jerking with as much force as we could muster. 264 Considerable variation in release efficacy was observed between trap models (Table 1), where 265 the Sauvageau traps generally released bears the least frequently, and Northwoods 155 released 266 bears the most. Finally, we used a come-along or a truck, with a scale affixed inline, to determine 267 the steady pull required to extract an adult foot from the Sauvageau traps. The average pull 268 required to free a foot was 164 kg (range=91-232 kg). Even after maximum pull was applied, we 269 did not detect superficial damage to toe bones or joints, although clearly significant pain would 270 be sustained by the bears during this time, and we are not certain that bears could always muster 271 the >230 kg pull required to ensure reliable release. 272 Based on the results of our experiment, we believe that better anchoring traps would 273 present safety concerns for bears, trappers, and wildlife professionals that might be called to 274 release bears. Although there is little work assessing the pull of a bear, a strong animal such as a 275 horse provides some insight into the forces animals can generate. A horse can exert ~70% of its 276 body weight at max exertion (Smith 1896). The adult female grizzly bear whose appendages we 277 used weighed approximately 140 kg, which would equate to a dead pull of ~98 kg if we use the 278 weight to exertion relationship from horses. This would be well below the maximum of >230 kg 279 required to reliably extract feet from these traps, and barely above the minimum recorded of 91 280 kg. As a result, bears may often be unable to escape from well anchored traps. There is a greater 281 chance that bears could free themselves if a long leash (>2 m) is used so bears could run and 282 generate more pull. It is thought that a 140 kg bear could generate between 400-1000 kg of pull 283 with a running start (Flaa et al. 2009 cons, and we hope that this work can help policy makers make an informed decision and choose 299 a solution that will resolve the amputated toe issue while ensuring trappers have sufficient 300 opportunity to trap furbearers. 301 Trappers have a long history of supporting conservation efforts and are sentinels of 302 changing landscapes and wildlife populations. However, trapping of furbearers in British 303 Columbia has received increased scrutiny amidst shifting sociopolitical trends in society. Bold 304 and novel strategies will be required from trapping organizations and wildlife management 305 agencies to maintain public support while also recruiting and retaining the next generation of 306 trappers. A tactful strategy to uphold public support is to proactively assess and react to 307 emerging evidence that threatens social acceptance of trapping (e.g., grizzly bears with 308 amputated toes), and where changes could help increase the well-being of animals. This 309 approach is the backbone of the North American Model, and while emerging evidence or science 310 can sometimes conclude no direct threat to wildlife populations, organizations must equally 311 weigh social ramifications and understand that the future of trapping may be predicated more on 312 social support than consensus over sustainability. Here we document a troubling issue where 313 grizzly bears are losing their toes after becoming stuck in baited furbearer traps. The timing of 314 furbearer seasons is the primary issue, with trapping seasons opening weeks before all bears have 315 denned, which creates a problematic period of overlap between active bears and baited traps. We 316 provide evidence of the issue, solutions to remedy it, and ultimately urge immediate action. 317

DATA AVAILABILITY 319
All analyses and supporting data can be found at https://github.com/ctlamb/Grizzly-MissingToes 320 321 322  represents one of the four unique individuals detected with amputated toes. Ci) an X-ray of the 425 foot of EVGM55, who was killed in a conflict with a cattle rancher. E) a body grip trap on the 426 foot of a hunter-killed grizzly bear from the Flathead Valley. F) the paw of a 1-year-old grizzly 427 bear stuck in a body grip trap captured by conservation officers near Pass Creek due to conflict 428 behaviour from a mom and offspring; significant infection was present, and the toes had nearly 429 completely separated from the foot by the time of capture. G) a grizzly bear with a body grip trap 430 on its foot detected feeding on a cow carcass by a remote camera west of Invermere. 431 Figure 3. Grizzly bears detected at trial marten body grip traps. The traps were wired so they 433 could spring but not close on bears' feet, and they were set for two weeks in areas frequented by 434 bears. Investigatory behaviour was common, and two of the four deployments were set off. 435