史維爾(Swales, 1981)的開拓性研究已經引發相當多人從事探討學術論文的結構問題。海仁德(Hyland, 2000)的研究顯示學術期刊論文的結構不同可能反映領域的文化和價值觀的差異。本文試圖探討這類結構不同也可能因論文作者的寫作經驗而有差別。我們是要觀察臺灣博士生所寫學術期刊論文「摘要」的結構特色。我們的語料分別取自五十篇本地生命科學系博士生期刊論文及三十篇相關領域的外籍專家論文。我們發現，幾乎所有八十篇語料含有「結果」與「結論」，可視為必用言步(obligatory moves)。就篇數而言，「背景」、「目的」及「方法」在博士生語料中的出現率分別為48%、52%和30%，遠低於外籍專家語料的80%、90%和60%。就篇幅而言，博士生花了高比例的篇幅(58%)去報導他們的研究結果，用20%的篇幅說明結論，用在其他言步者都少於10%。博士生作者認為論文摘有的主要目的就是呈現實驗結果。外籍專家比較兼顧其他各言步在論文摘要中的地位。最後，我們向臺灣博士生提出數點建議，或許能幫助他們改進摘要寫作。

　　　　　Seales' pioneering work (Swales, 1981) has inspired substantial research investigating the structure of academic papers. Hyland's (2000) study shows that structures of academic abstracts may reflect distinctive disciplinary cultures and ideologies. This paper attempts to show how such structures may also be influenced by the writer's writing experience. We aim to observe structural characteristics of academic journal abstracts written by Taiwanese Ph.D. students. Our data came from 50 journal articles written by local Ph.D. students majoring in life science and 30 from those written by foreign experts of the same field. We found that nearly all the 80 abstracts contain Results and Conclusion, the two obligatory moves in the corpora. The occurrence frequencies of Background, Purpose and Method in the 50 student corpus were, respectively, 48%, 52% and 30%, compared with 80%, 90%, 60% in the foreign expert corpus. The student abstracts spent 58% of text space on Results, 20% on Conclusions, and less than 10% on all the other three moves. The students authors appear to regard reporting results as their prime task in writing abstracts. Foreign scholars seem to put relatively proportional weights on the five moves. Our results might have some implications for teaching English for academic purposes in Taiwan.