摘要
台灣大多數山區地形崎嶇陡峭、植群密布，且食肉目動物大多生性隱密，造成直接觀察上的障礙。為了能夠提升食肉目動物調查效率，本研究自2013年10月至2014年3月，在行政院農業委員會特有生物研究保育中心中海拔試驗站以自動相機配合氣味站進行食肉目動物快速調查之可能性初探研究，於海拔1,600-1,800m 範圍中選定40個樣點架設紅外線自動相機(KeepGuard SD 1039)及氣味站(scent-station)進行調查。調查結果共記錄哺乳類5目11科15種，其中食肉目動物計4科7種，非食肉目動物中台灣野兔(Lepus sinensis formosanus)為本地區新記錄種。食肉目動物出現頻率(Occurrence Index, OI)方面，以黃喉貂(Martes flavigula chrysospila)(OI=13.0)最多、食蟹獴(Herpestes urva)(OI=4.5)及白鼻心(Paguma larvata taivana)(OI=4.4)居次。受到氣味站的吸引，黃喉貂、食蟹獴、白鼻心及華南鼬鼠(Mustela sibirica)等食肉目動物出現頻度較無氣味站自動相機的調查結果高，台灣黑熊(Ursus thibetanus formosanus)及鼬獾(Melogale moschata subaurantiaca)對氣味站沒特別反應。雖然從結果顯示氣味站的使用部分食肉目動物OI 值顯著提升，但我們無法排除是因為重複個體導致，使得OI 值無法用於估計物種豐富度，僅能表示該物種在此地活動頻繁，要了解食肉目族豐富度，往後的研究仌需要對個體進行標記後再進行族群數量的估計。食肉目動物多在氣味站架設後4~6天記錄到，建議相機的架設時間，宜滿足工作時數至少168小時。
並列摘要
Direct observation of carnivores is difficult in Taiwan due to the elusive nature of carnivores and the Island's dense vegetation coverage and many rugged mountains with steep slopes. As a way to effectively understand carnivores in Taiwan, we conducted a survey in Medium Altitude Experiment Station, Endemic Species Research Institute, COA, EY, from October 2013 to March 2014 via tested camera traps with scent-station. Forty plots were set up with 10 KeepGuard SD1039 digital camera traps and scent-stations. We recorded 15 kinds of mammal, among which Lepus sinensis formosanus is a new record in this area. Occurrence Index (OI) of species varied from numerous to common, which included Martes flavigula chrysospila (OI=13.0), Herpestes urva (OI=4.5) and Paguma larvata taivana (OI=4.4). In addition to Ursus thibetanus formosanus and Melogale moschata subaurantiaca, Martes flavigula chrysospila, Herpestes urva, Paguma larvata taivana and Mustela sibirica showed the highest OI. We also found Martes flavigula chrysospila kept Herpestes urva formosanus out from scent-station, and they were sympatric animals. The OI of Paguma larvata taivana and Viverricula indica taivana varied due to suspected food shortage in different seasons, which may cause sympatric carnivores to migrate. We recommend that camera traps should be set at least 168 hours to record almost all the mammals in the survey area.