臺灣出土古稻米粒的初步研究

Abstract

臺灣在北部、中部和南部都陸續有挖掘出碳化稻米的考古遺址；本論文試驗就是利用了其中的大坌坑文化南關里東遺址 (4800-4200 yr B.P.) 、牛稠子文化右先方遺址 (3800-3300 yr B.P.)、芝山巖文化芝山巖遺址 (3400-3000 yr B.P.)、蔦松文化五間厝遺址 (1400-1000 yr B.P.)、番仔園文化惠來里遺址 (1300 yr B.P.) 和番仔園文化南勢坑遺址 (600-500 yr B.P.) 等 ，共五個遺址所出土的碳化稻米作為試驗材料，從農藝學的角度去探討臺灣史前稻作農業的起源與演進。 由試驗結果推論，南關里東和右先方遺址因為具有相似的粒型，從育種的觀點比較兩遺址的種子大小變化趨勢，顯示兩遺址先民歷經1000年可能有的選拔稻種族群的粒型與粒徑的行為。與兩遺址出土的其他陶器特徵一樣，顯示二者之間可能有文化的傳承關係。 從這五個遺址在測量粒型結果的判斷，沒有發現偏野生稻的米粒粒型，推知臺灣約在五千年前，先民已將有秈、稉分化的野生稻馴化成為栽培稻，且因選拔適應當時環境栽種，出現了偏稉型的稻種族群，而隨著年代的漸近，可能也因為適應選拔的結果，逐漸有偏秈型的稻種族群出現。我們推論早期種植的稻種族群是由原生於臺灣本地野生稻所分化出來的可能性應該受到重視；而在蔦松文化五間厝遺址的古稻粒型分析顯示了爪哇稻之熱帶稉型稻的遺跡，可能是由於當時的海路興起，史前南島移民帶動了基因流，使得臺灣史前稻種族群的多樣性更加顯著。

The s up to about 5,000 years old had been excavated at several archaeological sites in the north, middle and south part of Taiwan. They are Tapenkeng Culture Nankungli East Site (NKLE；4800-4200 yr B.P.) , Niuchoutzu Culture Yihsuanfang Site (YHF；3800-3300 yr B.P.) , Chihshanyen Culture Chihshanyen Site (CSY；3400-3000 yr B.P.) , Niaosung Culture Wuchientso Site (WCT；1400-1000 yr B.P.) , Pantsaiyuen Culture Huilaili Site (HLL；1300 yr B.P.) and Pantsaiyuen Culture Nanshihkeng Site (NSK；600-500 yr B.P.) . The carbonized rice seeds from these archaeological sites were used as the materials to explore the origin and the evolutionary diversification process of landraces of rice, Oryza sativa subspecies Seng (indica) and O. sativa subspecies Keng (sinica / japonica), in Taiwan pre-historical agriculture. Based on the investigation, we found out that the same distribution of seed type and seed size in landrace populations between NKLE and YHF. The same fact is found in the characteristics of painted potteries excavated in these two archaeological sites, indicating that there might be a cultural connection between NKLE and YHF which located at the same place and are one thousand years apart. The changing pattern in both sites indicated that the rice landrace populations were selected for the seed size. By comparing the distribution of seed size, we did not find any wild perennial rice seed within each landrace population tested. We infer that the landrace populations were originated from local wild progenitor and had been domesticated. The distribution of Seng (indica) type and Keng (sinica / japonica) type within the population arose from two distinct gene pool of wild progenitor that diverged much earlier. The Keng (sinica / japonica) type-prone population (ex. NKLE；4800-4200 yr B.P. and YHF；3800-3300 yr B.P.) and Seng (indica) type-prone population (ex. WCT；1400-1000 yr B.P. ) might be resulted from adaptive selection following domestication in the middle and late Neolithic in these pre-historical areas. The unique large seed size found in WCT landrace population is identical to that of a tropical Keng (sinica / japonica) type, the now-a-day subspecies javanica. We propose that the gene flow of WCT landrace population might be changed by the pre-historical Austronesian immigrants in late Neolithic.