Baiting/Luring Improves Detection Probability and Species Identification—A Case Study of Mustelids with Camera Traps

Motion-triggered trail cameras (hereafter camera traps) are powerful tools which are increasingly used in biological research, especially for species inventories or the estimation of species activity. However, camera traps do not always reliably detect animal visits, as a target species might be too fast, too small, or too far away to trigger an image. Therefore, researchers often apply attractants, such as food or glandular scents, to increase the likelihood of capturing animals. Moreover, with attractants, individuals might remain in front of a camera trap for longer periods leading to a higher number of images and enhanced image quality, which in turn might aid in species identification. The current study compared how two commonly used attractants, bait (tuna) and glandular scent (mustelid mix), affected the detection and the number of images taken by camera traps compared to control camera sites with conventional camera traps. We used a before–after control group design, including a baseline. Attractants increased the probability of detecting the target species and number of images. Tuna experiments produced on average 7.25 times as many images per visit than control camera traps, and scent lures produced on average 18.7 times as many images per visit than the control traps.

Effects of scent lure on camera trap detections vary across mammalian predator and prey species

Camera traps are a unique survey tool used to monitor a wide variety of mammal species. Camera trap (CT) data can be used to estimate animal distribution, density, and behaviour. Attractants, such as scent lures, are often used in an effort to increase CT detections; however, the degree which the effects of attractants vary across species is not well understood. We investigated the effects of scent lure on mammal detections by comparing detection rates between 404 lured and 440 unlured CT stations sampled in Alberta, Canada over 120 day survey periods between February and August in 2015 and 2016. We used zero-inflated negative binomial generalized linear mixed models to test the effect of lure on detection rates for a) all mammals, b) six functional groups (all predator species, all prey, large carnivores, small carnivores, small mammals, ungulates), and c) four varied species of management interest (fisher, Pekania pennanti; gray wolf, Canis lupus; moose, Alces alces; and Richardson’s ground squirrel; Urocitellus richardsonii). Mammals were detected at 800 of the 844 CTs, with nearly equal numbers of total detections at CTs with (7110) and without (7530) lure, and variable effects of lure on groups and individual species. Scent lure significantly increased detections of predators as a group, including large and small carnivore sub-groups and fisher specifically, but not of gray wolf. There was no effect of scent lure on detections of prey species, including the small mammal and ungulate sub-groups and moose and Richardson’s ground squirrel specifically. We recommend that researchers explicitly consider the variable effects of scent lure on CT detections across species when designing, interpreting, or comparing multi-species surveys. Additional research is needed to further quantify variation in species responses to scent lures and other attractants, and to elucidate the effect of attractants on community-level inferences from camera trap surveys.

Evaluation of Novel Trapping Lures for Monitoring Exotic and Native Container-Inhabiting Aedes spp. (Diptera: Culicidae) Mosquitoes

Surveillance for diurnal container-inhabiting mosquitoes such as Aedes albopictus (Skuse), Aedes japonicus japonicus (Theobald), and Aedes triseriatus (Say) have routinely relied on the deployment of multiple trap types, including CO2-baited light traps, gravid traps, oviposition traps, and BG-Sentinel. These trap configurations have met with varying degrees of effectiveness and in many instances likely under-sample these key mosquito vectors. Most recently, the BG-Sentinel trap used in conjunction with the human-scent lure has been largely accepted as the gold-standard for monitoring Ae. albopictus. However, its ability to attract other container-inhabiting Aedes species has not been fully evaluated. During 2018, we tested new scent lures, TrapTech Lure-A and Lure-H (Bedoukian Research, Inc.), using BG-Sentinel traps with CO2 in two regions of Connecticut, Stamford and Hamden, against the BG-Lure. Pooled mosquitoes were additionally screened for arbovirus infection. A total of 47,734 mosquitoes representing 8 genera and 32 species were captured during the study, with the Stamford site deriving on average three times as many mosquitoes per trap, adjusting for sampling effort. Lure-A and Lure-H outperformed the BG-Lure in terms of total numbers, diversity evenness, and the proportion of both Ae. j. japonicus and Ae. triseriatus. There were no significant differences among lures in capturing Ae. albopictus, and in terms of species richness. Fifty-seven isolates of virus (West Nile, Jamestown Canyon, and La Crosse viruses) were obtained during the study, with no significant difference between trap-lure. We highlight both novel lures as effective attractants for use in mosquito surveillance=, which either outperform, or equal, BG-Lure.

Lack of Response to Olfactory Lures Among Mammals in Riparian Habitat in Southern Wisconsin

Non-invasive mammal surveys often employ olfactory stimuli on the assumption that they will attract mammals and increase the success of monitoring projects. However, information on the effectiveness of scent lures is variable and often relies on data generated from mammal tracks or sign, which can be challenging to quantify. Therefore, we sought to determine whether certain olfactory stimuli are more effective than others at eliciting a response from mammals along riparian corridors in Southern Wisconsin, using camera traps to monitor response to four scents and a control of de-ionized water. We recorded the number of times each species responded as well as the length of time spent investigating scents (Muskrat gland, mink gland, Red fox urine, crayfish oil) and the control. We recorded 2812 passes by 16 mammal species during our 12-month study, with Virginia Opossum (Didelphis virginiana), Woodchuck (Marmota monax), Eastern Gray Squirrel (Sciurus carolinensis), Eastern Fox Squirrel (Sciurus niger), Raccoon (Procyon lotor), Coyote (Canis latrans), and White-tailed Deer (Odocoileus virginianus)observed on cameras most frequently. However, 72% of observations involved mammals passing through without investigating any scents or the control, and among the 28% of observations where mammals did investigate, we found no significant differences in their response to specific scents or the control or in response by season. Further analysis revealed no significant differences in the time mammals spent investigating individual scents or the control. The lack of response suggests that factors other than scent may have attracted wildlife to our stations. Although under some circumstances olfactory attractants may increase the level of response to monitoring stations, we suggest that attraction to our stations was largely a behavioural response to novel stimuli in the environment.

Control of feral cats for nature conservation. III. Trapping

We present comparative success of various trapping methods trialed during control of feral cats at a site for the reintroduction of threatened mammals at Shark Bay, Western Australia. Our results come from 31 703 trap-nights that caught 263 cats (an average of 0.83 per 100 trap-nights). Cats differed markedly in their vulnerability to trapping depending on whether they primarily scavenged at rubbish tips or around human settlement or whether they hunted for their food in the bush. Cage traps were an effective means of controlling the former, with 9.4 cats captured per 100 trap-nights. Scavenging cats included a higher proportion of sub-adults and kittens and lower proportion of adult males than hunting cats. Variation between years in capture success for hunting cats was largely explained by the abundance of rabbits relative to that of cats and whether the rabbit population was increasing or decreasing. These factors accounted for a nine-fold difference in trap success. The number of cats caught in any particular trapping session could be explained by location (rubbish tip or town versus bush), trapping effort (typically greater effort yielded higher captures), abundance of cats at the site (captures were highest when cats were abundant), and season (captures were highest in the first half of the year when the young of the year were becoming independent). Concealed foot-hold traps, in a range of possible sets, provided effective methods for capturing cats that hunt, except where capture of non-target species was a critical limiting factor. Cage traps caught cats at a comparable rate to foot-hold traps for standard sets, but caught a significantly different cohort. Concealed foot-hold traps caught a higher percentage of adult cats, particularly males, than did cage traps. Mouse carcases and rabbit pieces were significantly more effective as lures when rabbits (the major food of cats at the site) were at low densities, whereas the success of commercial scent lures was unrelated to food availability. Significantly more cats than expected were caught using food as an attractant at times of food shortage (late summer, autumn and early winter) for both scavenging and hunting cats. In contrast, scent lures caught significantly more cats than expected in spring and summer when cats were defending access to mates and/or territory. Hence, no single trap type, trap set, or lure provided unequivocally superior performance over others. Control is likely to be best achieved by a variety of trapping methods and lure types used in combination, supplementing well timed poisoning efforts. Trap success is likely to be maximised by trapping at times when the dominant prey of cats are scarce relative to the number of cats and are decreasing in abundance.