RESEARCH ON ANCIENT RICE CULTIVATION IN DONGSHAN VILLAGE NEAR ZHANGJIAGANG CITY, JIANGSU PROVINCE, PR CHINA

WANG, Cailin, East Jiangsu Province Agricultural Academy of Science, and

DING, Jinlong, Suzhou Museum, Suzhou, Jiangsu Province, PR CHINA.

        The Dongshan Village sites are in E Dongshan Village 3 km S of Yangtze River and 18 km S of Zhangjiagang City, Jiangsu Province. They are on the E slope of Xiang Mountain (elev. 137 m), rising W and falling E and ca. 4 m above surrounding farmland. They measure 230x260 m in ca. 60,000 sq. m. In 1989, Suzhou Museum described a 2x5 m rice find in ditch 2. In 1990, a second excavation opened a 5x10 m rice part in the N site called T3, simultaneously opening a l0x10 m E rice find in ditch 1 (T4), for a total of 170 sq.m. Results show Majiabang and Songze cultural accumulations, the former primary, with thickest level >3 m. Dual cultural debris in grave 8, house 6, ashpit l, includes excavated pottery, jade carvings and ca. 300 stone tools.

        Excavation shows 8 levels in the N side of T3 section (Table 1). Under the plowzone are grayish white and blue-black levels 2 & 3 above red-oxidized levels 4 & 6, over blue-brown and dark yellow level 5 & 7 on dark ashy level 8. C14-dated levels, with level 2 in Songze period and levels 3-8 in Majiabang period, show absolute dates of 6000-8000 years.

        Recently, the lower Yangtze has recurring finds of Neolithic paddy rice, with Chiangnan Suzhou's Caoxieshan rice paddy sites 6000 years old (1) (7) (8), and Gaoyoulong Village paddy rice sites N of the Yangtze 7000 years old (6) (10) (11). East Dongshan Village sites are on the Chiangnan border, their earliest thick accumulation being 8000 years old.

        E Dongshan Village paddy rice extends new research on indica-japonica differentiation beyond earlier studies limited to carbonized rice. Despite thorough research on paddy rice origin using indica-japonica differentiation, carbonized rice use may have been fortuitous; i.e., findings show the error rate of its separation into indica & japonica approaches 39% (19). Thus, ancient paddy rice research has certain limitations.

        Recently, leaf blade phytoliths were used in ancient paddy rice origin and spread using indica-japonica differentiation (1) (3) (6) (7) (10) (12) (13). They are physico-chemically stable even under organism oxygenolysis, not decomposing but becoming part of the soil and retaining original shape. Their number is quite high, each g of leaf blade having several tens of thousands. They separate into indica & japonica phytoliths accurate to >85% (2) (4) (5). Neolithic pots not only involve mixing straw with clay containing phytoliths, but are important in archaeological dating. Therefore, archaeological age of soil and sherd phytoliths may be used to determine paddy rice cultivation through phytolith quantity and shape. Ancient paddy rice research on phytoliths does not lose its effectiveness. Xiao Jiayi et al. (14) once searched each level of T3 and T4 in E Dongshan Village sites for phytoliths, but not could separate indica & japonica. This research uses red oxidized soil and potsherds from there to separate cultivated rice into indica & japonica.

Material and Method

1. Analytical Provenience

        Comparative analytical specimens are 3 soil samples (T3 levels 3, 5 & 8), 5 red oxidized soil samples [(T3 levels 3-6 and level 3(2)] and 23 sherds (T3 levels 2 (1), 4 (4), 5 (3), 6 (2) & 7 (1), plus T4 levels 2 (1), 5 (2), 6 (2), 7 (3) & 8 (4). Except T3 level 2, T4 level 2 and plowzone, all are Majiabang culture (Table 2).

2. Analytical Method

(1) quantitative analysis

        Phytolith quantitative analysis was done on 3 soil samples (15) (16), treated according to Hiroshi Fujiwara's method. 1 g samples of dry soil were powdered and put in 15 ml glass jars. One quota (by wt.) purified glass sand was used as quantitative frame of reference, with water added plus dispersal agent dropwise and processing 20 min. @ 250w, 38 kHz ultrasound to eliminate phytolith surface adsorption by glutinous grain. Particle size rejection was done according to the rule of precipitation; <10µm small granule and other impurities. After final drying, samples were examined under SEM.

        Specific phytolith traits were used to distinguish plant origin; e.g.s, awns of paddy rice, reed, bamboo subfamily, barnyard grass, separately counted in same field of glass sand, with separate calculation for l g dry soil sample for each type and phytolith quantity (N) [p.89]

N=a x GW/SW x Np/NG

type: a=l g glass sand; e.g., this research uses 300,000 heavy 0.0256 g purified glass sand, therefore a = l/0.0256 x 3 x 10⁵

GW = glass sand wt. (g)

SW = soil sample wt. (g)

NP = phytolith number

NG = glass sand number

        In 1 g soil sample, the phytolith quantity N and each phytolith coefficient (corresponding to phytolith surface and plant dry wt./seed-wt.) is used to calculate the unit volume of soil for plant productivity. As this study uses phytoliths from a surface area of 1 m² of 1 cm thick soil, kg/m² x cm is equivalent to plant productivity. Here, the concept 'plant productivity' differs from normal annual plant productivity in being used only as comparative indices between different levels and their phytolith contents. Furthermore, the phytolith coefficient varies according to plant type and its cultivation. This study employs the 1995 coefficient obtained from 28 regional specimens in normal planted condition (8).

(2) sherd analysis

        Red oxidized soil and 23 sherds were analyzed (17) according to Hiroshi Fujiwara's method to withdraw 5 phytoliths. Take a suitable size sherd, clean with ultrasound, then remove moisture under vacuum. Use slight mechanical power to powder specimens, followed by ultrasound to precipitate <10μm granules and other impurities, and finally observing using SEM.

(3) shape analysis

        Cailin Wang et al.'s phytolith shape determination method (3)(6) was used on specimens, Table 2, a schematic of paddy rice leaf blade phytoliths. 50 rice phytoliths were stochastically chosen for each specimen, magnified x400 to determine length, width, thickness & b length. A,b separately expresses phytolith arc height and handle length, their ratios forming coefficient b/a. Each specimen had 3 determinations (total=150), the mean value expressing the phytolith trait value, and following each value according to 97 Asian paddy rice phytolith principal components analysis on indica-japonica differentiation (2).

Z4 = 0.497VL - 0.2994HL + 0.1357LL - 3.8154b/a - 8.9567

(Z4 < 0 for indica; Z4 >= 0 for japonica)

        VL, HL, LL and b/a are phytolith length, width, thickness and shape coefficient. Distinction results show ancient rice was distinguished by indica-japonica type.

Results

        1. 3 soil samples from T3 levels 3, 5 & 8 were examined for phytoliths (Photos 1-3). Table l of C14 measurements shows E Dongshan Village sites had rice for at least 8000 years.

        Table 3 shows quantitative phytolith results in each soil layer. Generally, 1 g soil has >5000 rice phytoliths, with level 3 approaching 80,000 and levels 5 & 8 >200,000. Converting to rice output, level 3 dried soil weight is 37 with 11 kg/m² - cm; levels 5 & 8 is 97-101 with 28-29 kg/m² - cm, showing paddy rice growth possibility is enormous in these soil layers.

Table 3: E Dongshan Village site T3 3 soil layer paddy rice phytolith quantitative results. Note: Unit kg/m2 - cm expression and 1m² 1 cm thick soil shows suitable plant productivity using phytoliths.

Table 4: E Dongshan Villages site T3 3 quantitative results of other phytoliths. Note: Unit kg/m² - cm expression and 1m² 1 cm thick soil shows suitable plant productivity using phytoliths.

        Table 4 shows other phytolith results. No barnyard grass phytoliths occur. Level 3 has mainly reed, then sandalgrass and bamboo phytoliths. Level 8 has few bamboo phytoliths. Level 3 has more bamboo and reed phytoliths.

2. Analysis of red oxidized soil and sherd phytoliths

        Table 5 shows analystical results of red oxidized soil and sherd phytoliths. Sample 2 soil has abundant paddy rice phytoliths but little of reed, bamboo, sandalgrass and barnyard grass. No phytoliths occur in the 4 other red oxidized soil samples. Sherd samples 7 & 18 have abundant paddy rice phytoliths. T3 3rd l paddy rice sample has many phytoliths (Photo 4). Converting, heavy rough rice output reaches 25 and 7 kg/m² - cm, while barnyard grass phytoliths expand, converting to 15 kg/m² - cm.

        Of 23 sherds, 13 have paddy rice phytoliths, 2 with very many in T5 5th (#7) and T4 7th (#18) (Photos 5-6), plus 12 with sandalgrass phytoliths. But reed, bamboo as well as barnyard grass phytoliths are in specimen 5. Overall, other phytolith contents are small.

3. Paddy rice phytolith shape analysis

        Phytolith length, width, thickness and shape coefficient were measured in 3 soil samples, 1 red oxidized soil and 2 sherds to determine indica-japonica differentiation, obtaining distinction values between l.15-3.62 (Table 6). Phytolith shape shows E Dongshan Village site japonica began 8000 years ago.

        Measurements were repeated in comparing East Dongshan Village and Caoxieshan sites with 26 japonica and 26 indica phytoliths. The 4 E Dongshan Village sites phytolith traits are all obviously bigger than japonica, their difference reaching 1% levels (Table 7) and much bigger than indica. The 4 E Dongshan Village sites phytolith traits exceed Majiabang period Caoxieshan, reaching 1% levels besides differet thicknesses; other trait differences are unremarkable (Table 7).

Table 6: E Dongshan Village sites paddy rice phytolith traits mean value and indica-japonica distinguishing values* where Z4=0.4947 x length -0.2994 x width +0.1357 x thickness -3.8154 x shape coefficient -8.9567 (when Z4<0 for indica and Z4>=0 for japonica)