|
Résumé :
|
Estimates of detectability and removal of carcasses have been determined for vertebrates but not for invertebrates, although the latter are more frequently killed on roads. The aim of this study was to estimate the detect ability and persistence of road-killed butterflies and to assess the relative contributions of the location on the road, volume of traffic, and presence of scavengers to the removal of dead butterflies. During three independent experiments, dead butterflies from 12 species were placed at randomly chosen locations along selected road sections (both on asphalt and on a verge) with varying vehicle traffic volume (including sections without vehicles) and scavenger abundance. Dead butterflies were counted, scavengers' behaviour was observed, and the traffic volume was measured. The detectability was higher for butterflies placed on asphalt (0.979) than on a verge (0.767) and increased with body size. The removal rate of dead butterflies from asphalt on roads with vehicles was four times faster (persistence probability of 12 h: 0.139) than on roads where vehicles were absent (persistence probability: 0.508). Overall, only 5% of butterflies persisted for a 48-h period. Vehicles and bird scavengers removed 14.9% and 9.6% of dead butterflies, respectively. The probability of dead butterfly removal by birds was positively correlated with traffic, indicating that the number of road-killed butterflies might be substantially underestimated on roads with high traffic. This shows that imperfect detection should be taken into account when estimating the number of road-killed butterflies, especially for butterflies on road verges. (C) 2016 Elsevier Ltd. All rights reserved.
|