A Low‐Cost Small‐Mammal Camera Trap for Research and Education

Use of camera traps has become common in many areas of wildlife ecology (Burton et al. 2015, Caravaggi et al. 2017, Agha et al. 2018, Wearn and Glover-Kapfer 2019). Despite their popularity for observations of medium-to-large animals, they are seldom used for observation of small cryptic animals such as mice or shrews. This is not surprising given that most of the cameras used were originally designed as "gamecams" for helping hunters identify fertile hunting locations, and few hunters are interested in species <1 kg in mass. Motion sensors capable of detecting small mammals at a distance would need to be so sensitive that their use would result in huge numbers of false detections caused by wind-shifted vegetation and changing shadows. Moreover, to avoid predators, small mammals are cryptic and typically make heavy use of available cover, so that they are difficult to photograph. Several recent attempts have been made to improve camera traps for detecting small mammals with "face-on" photography. The "selfie trap" in which a camera modified for close focus is built into the end of a box, tube, or tunnel has been used to take face-on photos of small mammals under snow (Soininen et al. 2015), detect and distinguish woodland small mammal species (Gracanin et al. 2019), compare rodent and shrew occurrence in different patches of a peatland landscape (Littlewood et al. 2021), document overlap and seasonal variation in diel activity of 10 species of arboreal and semi-arboreal mammal species (Gracanin and Mikac 2022a), estimate home range and movement patterns of four species of arboreal small mammals (Gracanin and Mikac 2022b), and monitor populations of two boreal vole species across a network of live-trapping grids (Kleiven et al. 2023). The "face-on" images were most useful for identifying taxa for which species distinctions are easily made from that perspective. In each of these cases the selfie trap design produced photos in which species could be identified and, in some cases, individual animals could be distinguished. For species reliably identified from a "top-down" perspective, McCleery et al. (2014) introduced the innovative "bucket trap" design, which used an inverted bucket fitted with a customized (for close focus) Reconyx wildlife camera. Holes in the sides of the bucket restricted entry to the camera field of view to animals the size of small mammals. The simplified background (Taillie et al. 2021) provided images of sufficient quality to identify species, and in some cases, individuals. However, the relatively high cost (estimated 400–500 per station) poses a challenge for wide application. We have adapted and simplified the McCleery et al. (2014) design to produce a "MouseCam" that provides top-down images of small-mammals suitable for detection of species and performing occupancy analyses (MacKenzie et al. 2017), for a cost of ~65 per station. The system uses a low-cost "mini trail camera" that features a 120-degree or wider field of view and low-glow infrared illumination and cost between 30 and 50 each (2024 prices), coupled with inexpensive 5-gallon paint buckets. The revised design includes features which reduce disturbance by raccoons and potential damage due to flooding and is resistant to human interference. Thus far, we have deployed MouseCams 314 times in coastal Virginia, with a median of 27 days per deployment for an aggregate of 9, 004 days of monitoring (Porter and Dueser 2024, Data S1). With a minimum of 1 min between photos, we captured a total of 52, 231 photos. Small mammals were featured in 75% of the photos and 14% recorded some sort of disturbance (trap tipped over, camera lens fogged, poor exposure), although only 7% of those disturbances were serious enough to make it impossible to identify a small mammal. We were able to identify 8 small mammal taxa during that trapping, typically to the species level (Fig. 1). Observed taxa have ranged in size from Sorex and Cryptotis shrews, which typically have mass <10 g to brown rats, which may have mass 500 g or more, with most observed species in the 15–80 g range. The system consists of a mini trail camera, two modified buckets, a single bucket lid, a small can or jar top and some wood or PVC pieces (Fig. 2). The inner bucket has two 4. 5-cm holes cut in the side of the bucket, near where the bails are attached, to provide a way for small mammals to enter or exit the bucket (Appendix S1, Video S1). To attach the camera, the bucket is inverted and the camera mounted through a hole cut into the top, and secured with a Velcro strap. Bait, if used, can be placed in the small can or jar top mounted to the inside of the bucket lid, along with two "walls of despair" which block raccoons and other mesopredators from being able to reach into the inner bucket. A camouflaged outer bucket is placed over the inner bucket to protect the camera from disturbance, and the inner and outer bucket secured together using a self-drilling hex-head screw so that they can't be separated except using a hex-head driver. The station can then be secured in place using stakes through the bucket handles/bails. Additional refinements include small holes in the lid and buckets to permit drainage (if tipped over a bucket with no holes can fill with rain and damage the camera) and an inspection hole in the inner-bucket top so that bait can be examined without needing to remove the camera. The use of bait is optional, but putting loose bait in the container secured near the center of the base of the camera trap provides a reason for animals to linger in the center of the camera frame. We've used cracked corn as bait because it cannot germinate if removed from the trap and has only a modest odor, so we are not luring distant animals. Sunflower seeds have also been used. If a more attractive bait is desired, hanging peanut butter wrapped in wax paper through the observation hole in the top of the trap has been effective, without drawing large numbers of ants into the trap, which occurs if the peanut butter is placed in the bait container directly. The market for inexpensive "mini trail cameras" changes rapidly. We have used Campark T-20, T-120 and Voopeak TC11 cameras, which turned out to be virtually identical, apart from branding. However, given the rapid turnover of model names, the focus here is on critical camera features, rather than specific models. Critical features are a wide-angle lens (120° or higher) and low-glow or no-glow infrared flash. The wide-angle lens allows most of the bucket bottom to be viewed from the distance of the bucket top and also has a closer focus than narrower-angle lenses. The infrared flash is less disturbing to animals than a white-light flash. Mini cameras also use 4 AA batteries instead of the 8 batteries used for larger trail cameras, reducing battery cost and camera size. Cameras are typically listed with ratings of 24-megapixels which is significant overkill given the quality of the lenses and lighting conditions. Setting the cameras to a reduced resolution of 3–5 megapixels is more than adequate. Battery lifetime depends on how many photos or videos are taken and how long the flash needs to stay on, but can be up to 5 months on a single set of batteries with moderate camera use. Setting the cameras to wait at least 1 min between photos reduces multiple photos of the same animal occupation of the trap, but longer times could also be used. However, inexpensive 32GB SD cards can hold at least 16, 000 5-megapixel JPEG images, so storage is not the issue. Most of these cameras also support capturing videos. These are more consumptive of both power and disk space and take longer to post-process. However, they can be valuable for educational purposes where it can be interesting to observe the animals for longer periods and for research purposes where you want additional angles and poses to help identify species. As Newey et al. (2015) note, low-cost consumer-grade cameras have their limitations in terms of image quality and reliability, but in our experience the ability to deploy 8–10 low-cost camera trapping stations for the cost of one "professional" wildlife camera often outweighs these disadvantages. As noted by Sikes et al. (2016), obtaining visual data using cameras is among the least intrusive ways of monitoring free-ranging mammals. In the small mammal context, two features help increase the safety of the camera trap for small mammals: dual access holes and "walls of despair. " Having two holes on opposite sides of the bucket assures that no animal will become trapped in the bucket with a potential predator, such as a snake (Fig. 3) or larger small mammal (Fig. 4). This arrangement also allows animals to exit the camera trap on the side away from any predator outside the camera trap. Second, the whimsically named "walls of despair" impede access to the interior of the bucket by mesopredators, in particular raccoons. Although the primary purpose is to keep raccoons from being able to access the bait container (Fig. 5), they also prevent mesopredators from being able to reach in to grab any small mammals currently in the trap (Fig. 6). Low-cost camera traps have utility for both research and education. For education purposes, even a single camera trap can provide interesting examples of what small mammals are found in various locations. Even suburban backyards or schoolyards can yield interesting photos of a surprising number and variety of small mammals. For research purposes, arrays of camera traps can be used to estimate small mammal diversity and occupancy (MacKenzie et al. 2017). MouseCams are not without their limits. Unlike traditional livetraps where animals can be tagged and released, typically images are not able to distinguish individuals for mark-recapture estimation of population density. Information on sex, mass, and reproductive condition are also unavailable. Similarly, species (especially congeners) that are similar to one another from a top-down perspective, may not be distinguishable using images alone. However, cameras that record the date, time, and temperature can yield information on the activity patterns of small mammals, and because the traps can accommodate multiple animals (Fig. 1d) at one time, some insights into the structure of small mammal social groups can be made that are not available from traditional live traps. This work was supported by NSF Grant DEB-1832221 to the University of Virginia. The authors have no financial interest or relationship with companies selling the components described in the paper. Photos of animals are currently available at http: //www. vcrlter. virginia. edu/mousecam and information about photos are available in the Environmental Data Initiative repository at: https: //doi. org/10. 6073/pasta/4fcbd3f981432a30e9ae8e8dd43e0a53. Appendix S1. Data S1. Video S1. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.