Abstract: Capercaillie, a large species of grouse, are sometimes killed when they fly into high-tensile deer fences. A fence design which is lower or has a less rigid top section than conventional designs would reduce bird deaths, but such fences would still have to be deer-proof. The short-term behavioural responses of farmed red deer (Cervus elaphus) to fences of five designs, including four that were designed to be less damaging to capercaillie, were measured. Five deer were located on one side of a fence with a larger group (20 animals), from which they had been recently separated, on the other. The efficacy of fences in preventing deer from the small group from rejoining the larger group was also recorded. In addition to a conventional deer fence (C) the four new designs were, an inverted “L” shape (L), a fence with offset electric wire (E), a double fence (D) and a fence with four webbing tapes above (W). Four replicate groups of deer were each tested for 3 days with each fence design. Deer paced the test fence line relatively frequently (a proportion of 0.09 scan observations overall) but significantly less when deer were separated by fences E or C compared to L, W or D (overall difference between fence types, P<0.001). Deer separated by fence E spent significantly more time pacing perimeter fences than deer separated by fences of other types (overall difference between fence types, P<0.01) but deer separated by fence C maintained a low level of fence pacing overall. Analysis of behaviour patterns across the first day and the 3 days of exposure suggested that the novelty of the test fences, rather than the designs per se, influenced the behaviour of the deer. Over the course of the study, no deer crossed either C or L. Three deer crossed E and two deer crossed both W and D. On this basis, field testing, particularly of fence L, would be a useful next step.