Ocelot home range, overlap and density: comparing radio telemetry with camera trapping A. Dillon à & M. J. Kelly Department of Fisheries and Wildlife, Virginia Tech, Blacksburg, VA, USA Keywords ocelot; Belize; camera traps; density; home range; MMDM. Correspondence Adam Dillon, Department of Fisheries and Wildlife, Virginia Tech, 147 Cheatham Hall, Blacksburg, VA 24061-0321, USA. Tel: +847 712 4582 Email: adillon@vt.edu à Current address: Adam Dillon, 2022 Dexter Ave N Apt B, Seattle, WA 98109, USA. Editor: Nigel Bennett Received 29 October 2007; revised 25 March 2008; accepted 27 March 2008 doi:10.1111/j.1469-7998.2008.00452.x Abstract Because ocelots Leopardus pardalis and other solitary carnivores are elusive and hard to study, little is known about their density and population status. In the past few years, camera trapping and mark–recapture statistics have been used to estimate the density of a number of felids. Although camera trapping is now providing baseline data for managers and conservationists alike, recent doubts have been raised concerning the accuracy of the standard camera trapping procedure. We used radio telemetry to gain new information on ocelot home-range size and spatial organization in Central America, and compared the radius of our average ocelot home range with the standard camera trapping buffer. We compared the resulting density estimates to assess the current camera trapping methodology’s ability to estimate animal density. Five adult ocelots (two male and three female) were tracked to determine an average ocelot home range of 26.09km 2 (95% fixed kernel) and 18.91 km 2 (100% minimum convex polygon), with males demonstrat- ing larger ranges than females. All ocelots had larger home ranges in the dry season. Male–male home-range overlap averaged 9% while female–female overlap aver- aged 21%. Males shared 56% of their range with a primary female and 16% with a second and third female, while females shared 58% of their home range with a primary male and 3% with a secondary male. Density estimates based on the average home-range radius (11.24–12.45 ocelots per 100km 2 ) were less than those determined from standard camera trapping methods (25.88 ocelots per 100km 2 ), but similar to those determined using twice the camera trapping buffer to estimate density (12.61 ocelots per 100 km 2 ). Our results suggest that a standard camera trapping protocol may overestimate ocelot density. Accurate representation of animal densities and standardization of density estimation techniques are para- mount for comparative analyses across sites and are vital for felid conservation. Introduction Historically, ocelots Leopardus pardalis ranged in large numbers from the southern United States to northern Argentina (Murray & Gardner, 1997). Owing to hunting pressure and habitat loss throughout the 1980s and mid- 1990s, ocelots were considered to be Vulnerable on the IUCN Red List (IUCN, 2006). Although bans on the international fur trade have decreased the ocelot’s status to a species of Least Concern (IUCN, 2006), habitat loss continues to threaten persistence and population data are scarce (Sunquist & Sunquist, 2002). Gaining information on status is difficult because, like many felids, ocelots are solitary and elusive by nature. Recently, remote camera trapping has been used to study a variety of felids (e.g. tigers Panthera tigris Karanth & Nichols, 1998; Karanth et al., 2006; jaguars Panthera onca Kelly, 2003; Maffei, Cuellar & Noss, 2004; Silver et al., 2004; pumas Puma concolor Kelly et al., 2008; ocelots L. pardalis Maffei et al., 2005; DiBitetti, Paviolo & DeAngelo, 2006; Dillon & Kelly, 2007; bobcats Lynx rufus Heilbrun et al., 2006; Kelly & Holub, 2008; Geoffroy’s cat Oncifelas geoffroyi Cuellar et al., 2006). The standard method to estimate animal density via camera traps is to estimate population size through cap- ture–recapture statistics and divide the abundance estimate by the effective trap area of the camera survey. The effective trap area is determined by placing a buffer, equal to 1/2 the mean maximum distance moved (1/2 MMDM) of all ‘recaptured’ animals, around the entire camera trapping grid (Karanth, 1995), or around each camera station (Silver et al., 2004). Because most camera studies lack data on the target animal’s home range, the 1/2 MMDM buffer is used as a proxy for home-range radius (Wilson & Anderson, 1985; Karanth & Nichols, 2002). Although camera trapping is quickly becoming an accepted technique for estimating felid abundance and density, recent studies have shown that reduced spacing between cameras (Dillon & Kelly, 2007), small survey area (Maffei & Noss, 2008) and lack of information on true home-range size (Soisalo & Cavalcanti, 2006) can Journal of Zoology Journal of Zoology 275 (2008) 391–398 c  2008 The Authors. Journal compilation c  2008 The Zoological Society of London 391 Journal of Zoology. Print ISSN 0952-8369