Thinking inside and outside the box: Comparing the efficacy of camera trap designs for surveying small mammals.

BACKGROUND: Small mammals are vital to ecosystem health and serve as key indicators of environmental changes. Population trends in the UK for many species have remained largely undocumented, however, emphasizing the need for more effective monitoring tools. This study evaluates the efficacy of three camera trap designs for detecting small mammals: no-box camera traps (NCTs), open box camera traps (OCTs), and closed tube camera traps (TCTs). METHODS: The study was conducted at Silent Valley Nature Reserve, South Wales, focusing on Fagus sylvatica woodland. The authors thank the reserve manager for providing verbal permission to access the site managed by the Gwent Wildlife Trust. Six baited survey stations, each equipped with one NCT, OCT, and TCT, were established in low-disturbance areas, placed approximately 5 m apart. Cameras recorded 15-second videos with a five minute trigger interval. Generalised additive models (GAMs) were used to assess temporal detection patterns, while hourly detection data and overlap metrics evaluated similarities between apparent activity patterns across trap types. RESULTS: Mice (i.e., wood Apodemus sylvaticus and yellow-necked A. flavicollis mice; n = 827) and bank voles (Myodes glareolus; n = 454) were the most commonly detected species in 1,874 detections of small mammals across 57 camera trap days. Both boxed designs yielded more detections of focal species and fewer false triggers than NCTs. TCTs recorded greater number of focal species detections relative to total detections (79%; compared to 58% for OCTs and 14.8% for NCTs) and the fewest unidentified species (16, compared to 51 and 49 respectively). Vole and shrew activity patterns differed between boxed and NCT designs. There was evidence of differences in detection among box designs, with shrews being more frequently detected by TCTs (n = 196), and voles more frequently detected by OCTs (n = 273). DISCUSSION: Boxed traps aided species identification but were affected by camera focal issues, which could be mitigated with low-cost adaptations. Our study highlights the need to consider focal species ecology when designing camera trap studies and provides a basis for the development of more effective small mammal surveys, survey methods and conservation tools.