Fig 1.
A) Parts of a MouseCam including the inner bucket (left), with 4.5 cm holes near the rim to allow animal access, and an opening in bottom of the bucket for the trail camera to be attached facing down, with a Velcro strap to secure it; the lid (lower right), equipped with a reservoir for bait, and “walls of despair” to frustrate predators attempting to access the bait; and the outer camouflage bucket (upper right). B) deployment of a MouseCam. The outer camouflage bucket provides protection for the camera within. It is secured to the inner bucket with a self-tapping hex screw. The bails of both buckets are attached to the ground using anchoring stakes.
Fig 2.
A) a hardware-cloth wire runway, B) a raccoon (Procyon lotor) attempts to reach a bait can, but is frustrated by “walls of despair”.
Table 1.
Comparison of detections from live trapping vs MouseCams.
Fig 3.
Example photos of different species taken in the forest plot and on Hog Island, VA.
A) white-footed mouse (Peromyscus leucopus, medium size, large eyes with strong light reflections, smooth pelage), B) marsh rice rat (Oryzomys palustris, large size, smaller eyes), C) house mouse (Mus musculus, small size, small eyes, lighter pelage), and D) brown rat (Rattus norvegicus, large size, rough pelage, thick tail, distinctive nose). The MouseCam in D did not include an attached bait container or any predator exclusion device.
Fig 4.
Results of the occupancy analysis for the forest plot, with 95% confidence limits for occupancy.
Percent occupancy is calculated as Ψ (from the model Ψ (.)p(.)) x 100. Diamonds indicate the naïve occupancy estimates.
Fig 5.
Probability of observation by door type, with 95% confidence limits.
Table 2.
Results of occupancy modeling.