從不同空間尺度探討達爾文歸化假說：以臺灣海岸地區為例

Abstract

達爾文早在1859年便提出外來種的歸化成功與否，會受到該引進地區是否有同屬物種存在的影響，因為同屬物種具有類似的環境需求，其競爭會比異屬物種間激烈，因此若該地區存在同屬原生種時，則不利於外來種歸化（此稱為達爾文歸化假說；Darwin’s naturalization hypothesis）。然而亦有學者認為，同屬物種通常偏好類似的環境，因此外來種也可能對有同屬物種建立的區域有較佳的適應能力（此一論述稱為環境篩選；environmental filtering）。在過去的研究當中，這兩種看似矛盾的假說則皆有研究支持，而這樣的結果可能是因為先前的研究未將空間尺度的效應納入考量。競爭排除的強度會因為空間尺度的增加而遞減，因此在測試達爾文歸化假說時，推測空間尺度對於結果的影響應該極為巨大，然而目前尚未有研究針對空間尺度的差異來探討達爾文的歸化假說。本研究調查在臺灣海岸地區，在不同空間尺度下（最小為1 m2；最大為縣市海岸範圍，平均約450 km2），歸化與原生植物的共存的情形。在本研究中，將分別從分類關係與物種譜系親緣距離這兩種觀點，來探討達爾文歸化假說在不同空間尺度下是否皆成立，以及造成此結果的機制為何。 首先，將檢驗各尺度下原生屬與外來屬歸化種之出現狀況，以及歸化物種與同屬原生種共存的情形。以分類關係表示物種相近程度時，使用了廣義線性混合模型（Generalized linear mixed model；GLMM）來分析各種空間尺度下，歸化物種出現的機率與同科或同屬原生種的物種數、以及歸化物種為原生屬或外來屬是否有關。以譜系親緣距離表示物種相近程度時，則計算歸化物種與整個原生群落的平均親緣距離、以及到最相近物種的親緣距離，用以檢視原生物種與歸化物種共存時的群落譜系結構為何。 結果顯示，雖然歸化物種外來屬的數目大於原生屬的數目，優勢的歸化植物多為原生屬。以GLMM的結果看來，歸化物種遇到越多的同科原生種時，無論在何種尺度下，歸化物種出現的機率都比較高。而同屬原生種的物種數在大尺度下為正向影響，在小尺度下卻會降低歸化物種出現的機率。此外，大尺度之下，原生屬歸化物種的出現機率較外來屬歸化種相對較低，顯示同屬原生物種亦會對歸化物種造成負向影響。檢驗不同尺度下的群落譜系結構時，則發現除了縣市尺度外，其他尺度皆為譜系聚集（phylogenetic clustering）。另外，以臺灣海岸整體看來，歸化物種與原生物種在譜系樹上重疊度不高，顯示兩者極不相似，符合達爾文歸化假說。 由這些結果看來，達爾文歸化假說不只在小尺度下成立，在大尺度下亦有可能。但是除了競爭以外，食植作用亦可能對與原生種親緣相近的歸化種造成壓力，因而抵銷了環境篩選的正向效果。在這些機制同時作用下，因為其在不同空間尺度的影響力不同，而產生了不一樣的結果。當競爭與食植的效應大於環境篩選的效應時，達爾文歸化假說即成立。然而譜系群落結構分析的結果與GLMM的結果不同，推測是由於譜系樹的解析度不足所造成。整體而言，在大尺度下，親緣相近的物種存在時，雖然可能代表歸化物種對該環境適應較佳，但亦可能有競爭、食植等負向作用存在，而且這些作用在小尺度下更為明顯。

Charles Darwin has proposed that exotic species would be less likely to naturalize if they have native congeners in the introduced regions, due to the intense competition with closely related species (Darwin’s naturalization hypothesis). However, since the existence of native congeners might also be a sign of better suitability for exotics in recipient communities, the contradictory view was also proposed. Furthermore, the effect of competitive exclusion between closely related organisms might diminish when the spatial scale increases. In this study, analyses were conducted at the scales from 1 m2 to county area (about 450 km2) along the coast of Taiwan to test Darwin’s naturalization hypothesis, and evaluated the relative importance of competitive exclusion and environmental filtering in determining the co-occurrence of natives and exotics, from both taxonomic and phylogenetic aspects. From the taxonomic aspect, the generalized linear mixed model (GLMM) was applied to test whether the probability of occurrence of naturalized species was affected by the presence of confamilial and congeneric native species at different spatial scales. From the phylogenetic aspect, to evaluate whether the naturalized species were phylogenetically distantly related to the co-occurred native species, the mean phylogenetic distance of naturalized species to the entire native species community (MPDN) and to their nearest native species (MDNN) were calculated at each scales. Standardized effect sizes of MPDN or MDNN were used to represent the overall phylogenetic structure under the specific scales. The findings showed that about one third of the plant species in the coast area were naturalized species. This proportion was much higher than that of the whole Taiwan area. Most of the naturalized species belonged to exotic genera, but most of the dominant naturalizers belonged to native genera. The occurrence of the naturalized species was positively associated with the number of native confamilial specie regardless of the spatial scales. The number of congeneric natives had negative effects at the small scales. The native generic naturalizers had relatively lower probability of occurrence at the large scales. Phylogenetic clustering was found at the scales of quadrats, habitats, and plots. For all species at the coast of Taiwan, the naturalized species were phylogenetically dissimilar to the natives, which corresponded with Darwin’s assertion. These results supported Darwin’s naturalization hypothesis at the small scales, and even some large scales, such as the county scale considered in this study. Both the competitive interaction and the herbivory effect could cause the occurrence of naturalized species negatively related to the presence of congeneric natives, masking the positive effect of environmental filtering. The mechanisms mentioned above operated simultaneously and their relative importance might change at the different scales. Darwin’s naturalization hypothesis would be supported if the effects of competition and herbivory overweighed the effect of environmental filtering. The outcomes of phylogenetic analyses contradicted the results from taxonomic perspective. This might result from the poor resolution of the phylogenetic tree. Overall, the naturalizers may only take advantage of similar preference of the environment with closely related natives at the large scale, but the negative effects of biotic resistance from their relatives could occur from the small to large scales.