台灣及金門、馬祖沿海文昌魚之系統分類及生態研究

Abstract

文昌魚屬於脊索動物門（Chordata）中的頭索動物亞門（Cephalochordata），演化地位恰介於無脊椎動物及脊椎動物間，且具有五億年前脊椎動物始祖的形態，因此具有高度生物演化上之意義。目前全世界對文昌魚的研究多半集中在高類元演化及組織結構上，本論文則是針對台灣及金門、馬祖沿海文昌魚之系統分類及生態進行研究。 台灣本島、金門、馬祖及台灣堆海域目前共發現有四種文昌魚，分別為鰓口文昌魚屬的白氏鰓口文昌魚（Branchiostoma belcheri），側殖文昌魚屬的短刀側殖文昌魚（Epigonichthys cultellus）、魯卡側殖文昌魚（E. lucayanus）及馬爾地夫側殖文昌魚（E. maldivensis），其中於台灣東北角所採獲之白氏鰓口文昌魚為台灣的新紀錄，魯卡及馬爾地夫側殖文昌魚則是首次於台灣北部海域發現。另外在文昌魚目中分類地位仍不確定的浮游型文昌魚（Amphioxides），本研究雖以DNA序列之比對證實於南灣後壁湖港外所採獲之四尾漂浮文昌魚（Amphioxides pelagicus）為魯卡側殖文昌魚之幼生，但仍不足以否定浮游文昌魚屬的存在性。 以較文昌魚為進化之八目鰻（lamprey）為外群，分析文昌魚粒線體12SrRNA片段所重建之類緣關係樹，顯示生殖腺及腹褶對稱發育的鰓口文昌魚屬較不對稱的側殖文昌魚屬為原始，以演化最儉約分析法更顯示鰓口文昌魚屬及側殖文昌魚屬皆為單系群，又因魯卡及馬爾地夫側殖文昌魚兩種側殖文昌魚間之遺傳距離遠大於與其他鰓口文昌魚之距離，且魯卡具有尾突此衍徵，馬爾地夫則無，此差異顯示側殖文昌魚應可再分為兩個不同的屬。又於金門、廈門採獲之白氏鰓口文昌魚與其在日本、青島之青島亞種（B. b. tsingtauense）在遺傳上並無分群或區隔之現象。 底砂之特徵為影響文昌魚底棲分佈的重要因素。白氏鰓口文昌魚、魯卡及馬爾地夫側殖文昌魚的地理分佈及出現頻度與底砂的粒度大小最相關。白氏鰓口文昌魚常見於金門料羅灣海域，多棲息於粗砂中，在本島及馬祖則極為少見；魯卡及馬爾地夫側殖文昌魚對棲地之選擇則無顯著差異，皆分佈於東北角及墾丁海域礁砂混合帶的中等粗砂中，但魯卡側殖文昌魚之密度遠高於馬爾地夫側殖文昌魚。本研究更發現兩者之棲地雖多重疊，但卻有不同之偏好，在顆粒較粗、泥量極少的環境中，馬爾地夫側殖文昌魚的相對密度大幅提高，而在細砂地上則多由魯卡側殖文昌魚所組成，顯示兩者在微棲地上的選擇有所區隔。 分析自1999年9月至2001年4月魯卡側殖文昌魚族群隨季節之變動，發現南部族群密度較高，全年皆有出現，於6月達高峰，無明顯之生殖季節，但6-10月生殖成熟之比例較高，應有新族群不斷在入添；北部族群密度則較低，僅於5-9月出現，數量也於6月達高峰，生殖成熟之個體則僅出現在水溫較高的8-9月。南北族群皆以濾食周圍碎屑維生，以矽藻為主。

Four species of lancelets, Branchiostoma belcheri, Epigonichthys cultellus, E. lucayanus and E. maldivensis were discovered from the coast of Taiwan, Kinmen and Matsu. B. belcheri collected from northeast coast was a new record of Taiwan, E. lucayanus and E. maldivensis were first found in northern coast of Taiwan. The taxonomic status of Amphioxides in Branchiostomatidae is not clear until now. It resembles the larvae stage of lancelets, but has larger size, more gill slits and even well developed gonads. In this study, we aligned partial sequence of 12SrRNA of mtDNA of A. pelagicus with other lancelets in Taiwan. And proved that this kind of lancelets is the larvae of E. lucayanus. The reconstructed phylogeny tree shows that Branchiostoma, which has symmetric gonads and metapleural, is more primitive than the asymmetric Epigonichthys. Branchiostoma and Epigonichthys are both monophyletic group in maximum parsimonious tree. But the genetic distance between E. lucayanus and E. maldivensis is much farer than Epigonichthys and Branchiostoma. Besides, the autapomorphic character, urostyloid process, is only present on E. lucayanus. According to all these differences, I proposed that two genus should be divided from Epigonichthys. B. belcheri collected from Japan, Tsingtau, Amoy and Kinmen had no differences in molecular analysis. The distribution and abundance pattern of lancelets is most related to the size of substratum. B. belcheri was most abundant in Kinmen which has coarse sand with some silt and clay but very rare in Taiwan. The habitats of E. lucayanus and E. maldivensis had no significant difference, both of them were found in coral reef with median sand, but the former was much more abundant. Although their habitats were overlapped, there was still some different preference between them. E. lucayanus was most abundant in fine sand, and E. maldivensis in very coarse sand without silt and clay. The dynamic pattern of E. lucayanus from September 1999 to April 2001 can be divided into two populations. The south population had higher density, appeared all the year, highest in June, had no obvious reproductive season, but the mature proportion was higher from June to October. I proposed that recruitment to south population happened all the year. The north population had lower density, appeared only from May to September, highest in June. Both populations fed on detritus, especially diatoms.