Adverse effects of hunting with hounds on participating animals and human bystanders

Hunting mammals with hounds is little studied. We present two datasets consisting of quantitative and qualitative data from self-selected respondents. The first came from hound handlers’ reports of hound injuries with post hoc verifications by government agents. The second came from by-standers reporting eyewitness encounters with hounds or handlers. Self-selected samples cannot be used to extrapolate rates in space or time but do provide nuances of human-animal and human-human interactions. From In the state of Wisconsin, USA, we describe government data on 176 hounds reported to have suffered injury during encounters with wolves. The government did not collect data on wolves or other non-target animals that may have been injured during these encounters. We investigate two wolf-centered hypotheses for wolf-hound interactions, find little support for either, and propose new hound-centered hypotheses. We also describe 105 human bystanders’ reports of experiences with hounds, handlers, and law enforcement agents.

200 invited any members of the public who described adverse hounding encounters to fill the online 201 report. Therefore, respondents were self-selected with attendant biases discussed below.

202
About 80% of respondents used the online form to report anonymously, and 20% sent 203 their responses directly to SCWC via mail, phone, email, or in person while being assured of 204 anonymity. SCWC administrators stripped identifying information from the data and shared the 205 data with LM for publication. We analyzed anonymized data stripped of identifying information. 211 Respondents could identify county in which the interaction occurred. Respondents could report 212 how many hounds they saw during each interaction. We used the average of the latter data to 213 cross-check with average hound party size from WHI records. When two respondents 214 mentioned the same interaction by date and county, but different numbers of hounds, LM 215 averaged and rounded up for the number of hounds. LM screened the sample to eliminate 216 responses that did not report an interaction but only expressed an opinion about the practice of 217 hunting with hounds (Supplementary Material 2). After the screening, the sample presented 218 here appears to come from independent incidents although we had no way to verify location or 219 date.

220
We evaluated two prior hypotheses for WHI: territoriality and predation. In brief, 221 territoriality might lead wolves or hounds to interact aggressively, which might be expected to 222 affect seasonal rates of WHI because either canid could have more to protect in certain 223 seasons. For example, wolf packs protect pups while young in summer or early fall. Hounds 224 have regular ranging areas and only part of the year access to those areas. Regarding the 225 predation hypothesis, periods of higher caloric need and lower food availability for wolves might 226 lead to higher rates of WHI in late summer or early fall (caloric needs) or winter (low availability).

236
We handled the problem of self-selection bias by discussing it in every section of this 237 paper. Also, we avoid extrapolating from the data but restrict analyses of percentages and 238 frequencies to within the dataset and note when upward or downward bias might arise from self-239 selection for extreme cases.

260
In total, 89% of WHI occurred while hunters reported pursuing black bears, bobcats Lynx 261 rufus 6%, coyotes C. latrans 4%, raccoons Procyon lotor 1%). However, we lack independent 262 data on the animal being pursued by those hounds at the time of WHI. We also lack the relative 263 frequencies of targeting each species with hounds statewide and over time. There was no 264 association between the outcome of WHI categorized as either injury or death and the prey 265 being pursued by hunters (X 2 = 1.9, p = 0.75, df = 4, n = 140). The bear-hound-training period 266 accounted for 62% of WHI, whereas the bear-hunting season accounted for 28% despite being 267 the same length approximately. Outcomes were not associated with month annually (X 2 = 8.5, p 268 = 0.38, df = 8, n = 176, Table 1). These two results suggest that seasonality and wolf 269 reproductive timing did not predict injury or death of hounds, which undermine both the 270 predation and territoriality hypotheses but see further below. We lack information on the 271 frequency with which WHI was initiated by hounds, which could result in different causal 272 hypotheses focused on the motivations of hounds and handlers.

286
Our analysis on hound body site bitten was limited to 109 WHI. We cannot be certain 287 that wolves inflicted every bite because of the above-mentioned delays in discovering hounds 288 and the potential bias created by compensation only for attacks by wolves. Of the 109 289 carcasses with bite information, 50 provided one bite location (46%), 37 provided two locations 290 (34%), and 22 provided 3 or more locations (20%). Considering all bite locations (n = 193), the 291 single most frequent bite site was the neck (33%), followed by back (17%), upper thigh (12%), 292 and chest (10%). We considered bites to the head, shoulders, neck (as opposed to throat), 293 back, and upper thighs as indicative the hound had been lower than its attacker. Those upper 294 body parts were represented in 72% of the 193 bites whereas under-parts (throat, groin, 295 sternum, ribs, lower legs, abdomen) were represented in 28% of bite locations. We found no 296 relationship between body site bitten and outcome, when we separated neck and head bites 297 from others (X 2 = 1.5, p = 0.22, df = 1, n = 66).

298
Another factor affecting the vulnerability of canids to attack by other canids is numerical 299 superiority. Aggression between wolves and coyotes in Yellowstone National Park had fatal 300 consequences when wolves outnumbered the smaller coyotes, but not when coyotes 301 outnumbered wolves, suggesting that group size exerted less influence than individual body Running head: adverse effects of hounding, p.14 of 27 336 greater height of wolves off the ground then off most hounds would predict such bite locations in 337 any case. In sum, we find equivocal support for both hypotheses. This could imply both are 338 correct or we are missing information, such as whether the hounds were pursuing Wolves. We 339 also do not have information on the body condition of any of the participants, which seems 340 essential data to test the predation hypothesis. 345 did not (64%). But we have no data on the use of bells among hounds that did not enter the 346 WHI database and the majority of records did not contain any information on devices. Outcomes 347 were not associated with hunter self-reports of affixing bells to collars. Another step handlers 348 might take to protect hounds and wolves would be to release hounds only in low-risk areas.
349 The state did not systematically collect data on whether their warnings about high-risk WCAs 350 were heeded. The WHI records did not contain such information.

368
Regarding bystander-handler interactions, 31% reported threatening altercations with 369 handlers, including being unwillingly detained by handlers 'trucks on public roads, or their own 370 private driveways. Of the 105, 51% of respondents reported they "feel intimidated by hound 371 handlers," and 44% feared retaliation from handlers for reporting confrontations to law 372 enforcement.

373
Finally, approximately a third of respondents described distrust of the information 374 provided by law enforcement officers and also reported filing official complaints upon which no 375 discernible action was taken. Of the 105, 36% believed a conflict of interest existed for law 376 enforcement officers, including game wardens, because of relationships between officers and 377 handlers, or because the officers were believed to hunt with hounds themselves.
378 Comparing numbers of hounds from WHI and survey data.

379
Survey respondents reported 2-8 hounds per interaction (average 3.7, mode 2). That 380 average is identical to the average number of hounds that handlers reported in their pack in 381 WHI above despite the different study periods and presumably locations. This seems to be 382 corroborating evidence of accuracy in reports of hound pack sizes in both datasets, as neither 383 set of complainants was aware of the other. Given the rarity of single hounds (5%) in WHI, the 384 bystander reports of >1 hound seem unsurprising. Similarly, bystanders reported >6 hounds in 3 385 events (8% of reports that include these data) but handlers never reported >6 in their pack after 386 a WHI. The legal limit per handler was ≤6 hounds per handler released from direct control.