2277Fax -+61 9360 6303 2 Highlights Roaming behaviour was unchanged when cats wore predation deterrents. Deterrent effectiveness stems from interference with hunting success, not less roaming. Housing density was the strongest predictor of home range and roaming extent.
ABSTRACTRoaming pet cats kill and harass wildlife, hybridise with wild felids, interbreed with feral populations, spread disease or annoy neighbours, and endanger their own welfare by fighting, being struck by vehicles or ingesting poisons. Confinement of pet cats is unpopular, so alternative methods to curb roaming behaviour would benefit wildlife conservation and pet wellbeing. Some owners whose cats participated in previous trials testing the effectiveness of the collar-mounted predation deterrents the CatBib and the Birdsbesafe collar cover (BBS) in reducing predation by pet cats reported that their cats stayed closer to home when wearing the devices. Therefore we tested whether these devices might curb roaming behaviour of pet cats as an alternative to confinement.Thirty cats participated. Trials occurred in spring and autumn in Perth, Western Australia (southern hemisphere spring -autumn). Cats wore GPS collars for 10 consecutive days, wearing the GPS collar alone for five days and wearing either a CatBib (16 cats) or BBS (14 cats) as well for a further five days. Treatment order was determined randomly for each cat. We represented cats' home ranges with 95 % kernel density estimates (KDE) (100 % minimum convex polygon (MCP) provided for comparison with other studies) and 50 % KDE (core home range). We also used data for all cats when not wearing either predation deterrent, plus data on a further four cats, to determine the relative effect of sex, age, night confinement, housing density, number of days of rain, total rainfall, and mean maximum temperature on both estimates of home range size.Neither device reduced home range significantly. The mean home range (95 % KDE) was 2.79 ha with the CatBib and 2.46 ha without. Figures for the core home range (50 % KDE) were 0.63 ha and 0.71 ha respectively. The mean home range (95 % KDE) with the BBS (where the sample included fewer cats from lower housing densities) was 0.58 ha and 0.50 ha without. The means for the core home range (50 % KDE) were 0.15 ha and 0.14 ha respectively. When cats were not wearing either device, 95 % and 50 % KDE were predicted most strongly by housing density, presumably a surrogate for cat density.Owners may use a CatBib or BBS to curtail their cat's hunting behaviour, but curtailing roaming behaviour needs another solution. Confinement, although unpopular, remains the most effective option where unwanted roaming is a problem.