R Wu, « An integrated study of Neolithic wetland agroecology in the Lower Yangtze River basin, China », Oxford Research Archive, ID : 10670/1.a792d2...
The Lower Yangtze River basin (LYRB) is one of a number of important centres of agricultural origins worldwide. The agriculture that emerged here was mainly focused on rice (Oryza sativa L.), particularly paddy rice. Previous research showed that there were indications for the occurrence of dryland components in the wetland agroecology of the LYRB. In addition, the ‘unexpected’ peak of the ratio of domesticated-type rice spikelet bases around 9000 BP suggested the possibility of multiple rice domestication in this region. Last but not the least, the variation in wetness and field size/structure suggested the co-existence of various regimes of rice management within or between geographic regions in the LYRB. All of this suggests that the framework of the formation of rice agriculture in the LYRB has yet to be built. Methods that can address the ‘how’ questions of the formation process of rice agriculture are required, i.e. those from an agroecological perspective. In this study, such perspective was taken to shed light on the growing conditions of rice (such as wetness, soil fertility and disturbance) and on the interaction between humans and non-cereal food plants, using water chestnuts as a case study. This study showed that stable carbon isotope values of rice grains are useful in differentiating modern paddy rice and dry (rainfed) rice, since paddy rice δ13C values tend to be lower, reflecting higher water availability and a corresponding shift in stomatal conductance. Paddy conditions also tend to increase the δ15N values of rice grains, suggesting that relatively high δ15N values (for example, 4.5 per mill to 8 per mill, as presented by the control group from fields in Shanghai) are caused by denitrification. In addition, several batches of samples from China and South Korea showed a clear negative correlation of δ13C values and δ15N values, which could suggest that this effect is exaggerated with manuring. The effect of intensive manuring may not be detectable in the initial growing seasons of application as increase in δ15N values was observed in glasshouse pot experiment samples but not in field samples. Functional weed ecology was demonstrated to be useful in classifying the intensity of modern paddy rice plots. Also, a wetness model was developed which successfully separated paddy plots in Japan from dry fields in Europe as well as one paddy plot under fallow. This model was then applied to weed assemblages from paddy fields (Late Liangzhu culture phase) at Maoshan, and the result was in line with previous archaeobotanical studies, suggesting a high wetness level realized by large amount of labour input. The combination of stable isotope analysis and functional weed ecology is a promising avenue for studying local arable farming conditions, the multi-threaded formation process of rice agriculture in the LYRB and the spread of rice farming in and outside China. Size and shape suggested that Tianluoshan water chestnut were not purely wild species. However, the ratios of morphological measurements which were useful for differentiating modern wild and managed water chestnut did not clearly characterise Tianluoshan specimens. A model using morphology of the detachment point of water chestnut fruits to indicate whether they were unripe or biologically ripe was developed to assess how and when the fruits were harvested, as one aspect of management. Biologically ripe water chestnut specimens appeared from phase 8 to beneath phase 6, which belonged to the first stage of occupation at Tianluoshan. These specimens extended the seasonality of water chestnut from late summer/autumn to winter. Collecting them in winter revealed the flexible role that this aquatic starchy species played in the subsistence economy.