恆春山茶之族群內遺傳結構

Abstract

恆春山茶(Camellia hengchunensis Chang)為台灣特有種，分布僅限於恆春半島南仁山附近山區之衝風坡面或稜線上。依據記載與觀察，山茶屬植物的傳粉媒介多以昆蟲(蜂類)為主，而種子由重力傳播。本研究目的在以微衛星分子標誌探討恆春山茶的交配系統(mating system)、種子與花粉傳播能力、族群之空間分布類型與遺傳歧異度、遺傳結構，並結合以上各點探討兩個假設：(1)小尺度下的遺傳結構顯著但不強；(2)小徑級的植株間遺傳結構較明顯，隨徑級增加，遺傳結構漸不明顯。對欖仁溪樣區141個樣本與南仁湖樣區12個樣本進行親子分析的結果，推測恆春山茶異交的比例相當高。種子在欖仁溪樣區內平均傳播距離為44.75公尺，最小值0.286公尺，最大值252.672公尺，但大多數案例在20公尺內。花粉在樣區內平均傳播距離為96.371公尺，最小值0公尺(自交)，最大值265.798公尺。遺傳歧異度方面，Shannon’s information index值為0.2712，Nei’s gene diversity值為0.1587，欖仁溪樣區內兩山頭間之Gst值為0.0236，Nm值20.6892；欖仁溪樣區與南仁湖樣區之樣本間Gst值為0.0761，Nm值為6.0734。樣區內所有植株無論在任何空間尺度下均呈現聚集空間分布，而遺傳結構在空間距離40~60公尺以內較為顯著。將所有植株區分為三個徑級後，最小的徑級(小樹)與最大的徑級(老樹)在小空間尺度下皆呈現隨機空間分布，其遺傳結構亦大致落入隨機區間。成樹在任何空間尺度下均呈聚集空間分布，遺傳結構在空間距離30公尺以內顯著。以上結果意味著恆春山茶的種子主要是重力傳播，但可能有其他未知的傳播機制。其花粉傳播距離則可遠達一公里以上，使樣區之間的遺傳分化不明顯。所有植株在小尺度下遺傳結構顯著但不強，符合本研究的假設(1)。而此現象可能受自交比例低，基因交流受限，成株密度高且族群結構呈反J型的共同影響所致。但假設(2)在恆春山茶中並不成立，推測可能在成樹建立時期因為(a)老樹產生成樹這個世代時受某些因素影響，僅部分植株產生後代。(b)棲地異質性使某些家庭的子代存活率較高，使遺傳結構呈現出家庭成員的分布範圍。

Camellia hengchunensis Chang is an endemic species of Taiwan, which is distributed only on the windward slopes and ridges of Nanjenshan area in the southernmost Taiwan. According to previous research and observation, Camellia is pollinated by insect, often by bees, and the seeds are dispersed mainly by gravity. The object of this study is to use microsatellite markers to help us understand more about the mating system, seed and pollen dispersal, genetic diversity and spatial distribution pattern of C. hengchunensis, and to test two hypothesis about intra-population genetic structure of C. hengchunensis: (1) The genetic structure would be significant but weak at small scale; and (2) Genetic structure would decrease as the size class increases. The results of parentage analysis on 141 samples from Lanjenshi plot and 12 samples near Nanjenlake plot suggest that the outcrossing rate of C. hengchunensis is high. The average seed dispersal distance in Lanjenshi plot is 44.75 m, with a minimum value of 0.286 m and a maximum value of 252.672 m. The average pollen dispersal distance in Lanjenshi plot is 96.371m, with a minimum value of 0 m (when inbreeding) and a maximum value of 265.798 m. Values of Shannon’s information index and Nei’s gene diversity are 0.2712 and 0.1587 respectively in Lanjenshi plot. The Gst value of the two subpopulations in Lanjenshi plot is 0.0236, and the Nm value is 20.6892; the Gst value of Lanjenshi plot and Nanjenlake plot is 0.0761, and the Nm value is 6.0734. All samples in Lanjenshi plot form a clump spatial distribution at all scales, and the genetic structure is significant below 45~60 m. According to minimum DBH for onset of flowering, we divide all samples into three size classes: juveniles, adults and old trees. Both juveniles and old trees are randomly distributed at small scale and show a clump pattern at larger scales in Lanjenshi plot, and the genetic structure of them are not significant at all scales. Adults are aggregated at all scales, and the genetic structure of these individuals is significant below 30 m. The results presented above indicate that seeds of C. hengchunensis are mainly dispersed by gravity, but there might still be other dispersal mechanisms. The pollen dispersal distance can be over 1 km, which makes the genetic divergence between plots not significant. The first hypothesis about genetic structure is supported, which may result from low selfing rate, restricted gene flow, high density of adults and the inverse J-shape population structure. The second hypothesis about genetic structure is not supported, and the possible reasons might be: (1) For unknown reason, only part of “old trees” born “adults”; and (2) Only adults belong to some family survived during the establishing process, because of the heterogeneity of habitat.