本文將2002－2004年巴丹考古計畫的放射性斷年結果，與呂宋島北部類似的資料予以分析比對。由於礁灘貝與陶器破片斷年被認為不若木炭標本可靠，根據木炭標本的年代，顯示出巴丹島的移民是在相對來說較晚的西元前800年。木炭斷年把原本早期的Sunget和Naidi階段合併成一個時期。由於呂宋島北部的新石器遺址較古老（西元前1700年，木炭斷年），早期從台灣出發的新石器文化遷移可能繞過了巴丹島。早期巴丹物質文化顯示出與台灣和呂宋島北部的相似性，而語言學的資料無法斷定保守的巴丹語是反映出它比呂宋島北部的移民早，或者是巴丹島的與世隔絕。我們對東南亞地區島嶼的地理環境，以及晚期全新世島嶼移民類型的考量，並不支持新石器文化的傳布是跨島而行的單一路線。至少應有兩階段的傳布；也就是假設有一個較早的西南地區及較晚的東北地區。這種假設導引出晚期全新世這個地區的自然環境和互動情形，並指向新石器文化的「目錄」可能可以再分割更細，各文化間的關係比想像中的複雜。由於有關農業擴張情形的研究相對稀少，它一般所造成的人口成長，還不被納入新石器文化快速擴張的因素。特別是從紀元前2000年前之航海技術演進和氣候變遷來看，關於全新世晚期東南亞海島地區的人類移動，還可以有其他進一步探討的途徑。

Radiocarbon dating results from the Batanes archaeological project (2002-2004) are analysed and compared with similar data from northern Luzon. It is argued that the reef-shell and pottery residue ages are not as reliable as dates on charcoal, and that the latter show a relatively late colonization of the Batanes, beginning about 800 BC on charcoal ages. The two successive early phases originally proposed, Sunget and Naidi, are amalgamated into a single early period. As early neolithic sites in northern Luzon are significantly older, beginning about 1700 BC on charcoal dates, early neolithic movement from Taiwan might have bypassed the Batanes. Early Batanes material culture exhibits similarities with types and assemblages in both Taiwan and northern Luzon, and linguistic data cannot determine whether the conservative Batanic languages reflect colonisation earlier than settlement of Luzon, or isolation in the Batanes. Consideration of the geography of island Southeast Asia and of the general pattern of late Holocene island colonization does not offer support to simple steppingstone models of neolithic dispersal. At least two phases of dispersal, one earlier from the southwest, and one later from the northeast, can be hypothesized. Their potential existence raises questions about the nature and interaction of late Holocene populations in the region, suggesting that neolithic cultural inventories may have been more partitioned, and relations more reticulate, than is sometimes envisaged. As expansion of farming is relatively poorly documented, its common demographic impact probably does not account for rapid neolithic expansion. The evolution of maritime technology and climatic change, especially around 2000 BC, suggest alternative avenues for further investigation of late Holocene movements in island Southeast Asia.