鹿科動物因需多樣化棲地環境與食物資源供其生存，其對苗木啃食與植群演替的影響一直是森林經營上關注的焦點。自相關單位在墾丁地區復育並野放臺灣梅花鹿(Cervus nippon taiouanus)以來，目前在高位珊瑚礁林中已可發現其蹤跡。為了解梅花鹿對不同森林生育地之使用情形和季節差異，與其對林下苗木的啃食情形，及比較有無鹿隻出現對苗木生長差異，於2005年9月至2006年7月在林業試驗所恆春研究中心之喀斯勒森林永久樣區(10ha)，依樣區中四類生育地（紅柴型、茄苳型、蟲屎型和皮孫木型）之比例隨機選取60小區(10x10m2)，每兩個月調查一次各小區中梅花鹿排遺量與記錄林下苗木被啃食、折枝與死亡情形。並分乾濕兩季（2006年1月與7月）對全區1000小區進行排遺量調查，於乾季時調查各小區之地被層與珊瑚礁等棲地環境因子，以分析上述環境因子及林木數目與乾季有無梅花鹿排遺的關係。另在永久樣區外設立四組圍籬試驗樣站，以架設圍籬(實驗組)和未設圍籬(對照組)方式比較苗木生長情形。結果顯示乾濕季梅花鹿在全區皆呈聚集分佈，兩季在排遺量等級與有無排遺小區數上皆有顯著差異，在四類生育地有無排遺小區數量檢驗達顯著，但各生育地小區總數與有排遺小區數檢驗則不顯著，且有排遺小區數在兩季間無顯著差異，此表示梅花鹿在兩季對生育地使用依供應量之相對數量分配，無偏好性。以有排遺小區數計算兩季重疊度(OIA)為0.306，且四類生育地的重疊度相近(29~31%)。依各小區排遺量計算的兩季Pianka重疊度(OIB)為0.571，以蟲屎型貢獻40%最高。四類生育地中各小區排遺量在乾濕兩季間有顯著差異，且以茄苳型與蟲屎型兩季間的差異達顯著。棲地環境分析顯示生育地類型、林木數量、珊瑚礁岩覆蓋、草本和蕨類覆蓋等因子對乾季梅花鹿的出現有顯著作用。對60小區的分析顯示，有排遺小區數和排遺量皆在1月最高，相鄰兩次調查間OIB值在2006年1~3月間最高，以皮孫木型生育地貢獻比例較高，六次調查中月份各小區的排遺量在生育地間有顯著差異。梅花鹿啃食苗木有季節性變化，以2006年1月的啃食量最高。鹿隻大多取食高度在60公分以下苗木，苗木被折枝量不多，對此區優勢種黃心柿的啃食數量最多，但造成的死亡率與未受啃食相距不大。圍籬試驗樣站苗木生長以實驗組略優於對照組，且死亡率較低。由結果推論，目前研究地內梅花鹿密度不高，對苗木啃食仍未造成嚴重危害，但仍需持續注意其對林木更新與植群演替的可能影響
Cervidae lives on diverse habitats and food resources. Impacts of deer browsing on the survival of saplings and the succession of vegetations have been among the major concerns in forest management. Formosan sika deers(Cervus nippon taiouanus) has been restored and released to wildlife by Kenting National Park. At present , deers can be found in the karst forest of nature reserve. This study is aimed to examine the habitat use pattern and sapling browsing by sika deers in different seasons, and to compare the growth of saplings with and without the appearance of deers. 60 sub-plots(10x10m2) were chosen randomly in proportion to the amounts of the four habitat types ( Aglaia formosana type, Bischofia javanica type, Melanolepis multiglandulosa type,and Pisonia umbellifera type) in the karst forestry permanent plot(10 ha) managed by Hengchun Research Center, Taiwan Forestry Research Institute. From September 2005 to July 2006, a 2x2m2 sapling plot was set in each sub-plot , all the saplings within the plot were marked in September 2005. the sub-plots were surveyed bimonthly for feces and the marked saplings were monitored for browsing evidence and survival. Fecal counts were conducted for all 1000 sub-plots in dry and wet season (January and July 2006, respectively). To examine the relationship between habitat characteristics and the appearance of deer faeces, understorey vegetation and karst coverages were recorded for each sub-plot in dry season. Four pairs of fenced and unfenced plot were set outside the permanent plot to compare the growth of saplings. The results reveal that Sika deers tend to aggregate in both dry and wet seasons, faeces within plots differed significantly between dry and wet season, but they show no preference for the four habitat types. The overlap index(OIA) of plot with faeces is 0.306 between dry and wet season. Pianka overlap index(OIB) is 0.571, 40% of which was by Melanolepis multiglandulosa type. The amount of faeces within plot of the four habitat types differed between dry and wet seasons, especially in Bischofia javanica type and Melanolepis multiglandulosa type. Logistic regression analysis shows habitat types, numbers of trees, karst herb and fern coverages have significant effects on sika deer presence in dry season. In the six bimonthly surveys, the number of plots with feces and the mount of faeces within the 60 sample plots were highest in January. In addition, OIB is the greatest between January and March 2006 with the greatest contribution from Pisonia umbellifera type. There were also significant differences in amount of facees in the four habitat types. The proportion of the browsed saplings was the greatest at January 2006. Browsing by deer was found mostly on saplings under 60cm, and the leading shoot broken by deer was uncommon. More saplings of Diospyros maritime, the most dominant species in the karst forest, were browsed by sika deer in the study plots, but the mortality of the browsed saplings was similar to that of the un-browsed saplings. The growth and survival of saplings in the fence of plots were better than those in the unfenced plots. According to the result of the study, the density of sika deer and effects of browsing on saplings are not high by now, and the impacts on forest regeneration and succession need furthur attention.