(translated by Bing Shi; edited by George Leir and Bryan Gordon)

        The rise of agriculture is the basis of the incipient Neolithic, while the beginning of rice farming is an important part of the origin of agriculture. Neolithic archaeological data and excavation in the lower Yangtze Valley indicate the growth of rice farming is divisible into Domestication, Incipient and Developmental stages (Table 1).

A) Yuchan Cave


        Using published archaeological material, we believe that Chinese paddy rice agriculture began in the Yangtze Valley. Archaeological sites of the Neolithic Domestication stage are Xianrendong (10,000 year-old Fairy Cave) and Diaotonghuan (1) (Bucket Handle Cave) in Wannian County, Jiangxi province, and Yuchan Cave (Jade Striped Toad Cave) in Tao County, Hunan province. Xianrendong and Diaotonghuan sites are on a west branch of the lower Yangtze, while Yuchan Cave is in a relatively low area of the Nanling northern foothills (Fig. 1). From September to November of 1995, the Beijing University Archaeology Department divided the Xianrendong excavation into an Upper Accumulated sequence (east area with 2 levels, west area with 2 & 3 levels) and a Lower Accumulated sequence (3 and 4 levels). Findings from Upper and Lower sequences were very different; e.g., sandy clay sherds were in the Upper sequence but not the Lower. While there were sea mollusks and univalve shells in the Upper sequence, they were scarce to nonexistent in the Lower. There were ground stone tools in the Upper sequence, but only flaked tools in the Lower, mostly quartz, but also many small flaked flint tools.


click to enlarge

        Diaotonghuan is a cave remnant on a 30 m ridge about 800 m from Xianrendong. Data from excavated material suggest it was a temporary hunting camp or butchering area for Xianrendong inhabitants. Large quantities of bone, partly ground stone tools, bone and perforated shell, and coarse sandy clay sherds were in the upper levels, while lower level materials resembled those in the Xianrendong Lower sequence.

        Analysis of Xianrendong and Diaotonghuan excavated material suggested that the Upper and Lower sequences were 14,000 year-old Early Neolithic and 15-20,000 year-old Late Palaeolithic.

        Pollen analysis showed a sudden increase of grain vegetation, with pollen grains much bigger, like paddy rice. Phytolith analysis showed that Upper sequence grains were fan-shaped like paddy rice, providing important clues to the origin of paddy cultivation.

a) Yuchan Cave

        In 1993 and 1995, two excavations were carried out at Yuchan Cave (2) in Bai Zai Chun (White Stronghold Village), in Zao Yen Zheng (Long Life Goose Town), Tao County, Hunan. Stratigraphic thickness was 1.2-1.8 m. Except for signs of modern grave disturbance, cultural remains consisted of simple chipped cores, flakes, choppers, axes, scrapers or planes, knives, and hoe-like tools with odd curved digging handles. Important discoveries were primitive black-brown, poorly fired, 2 cm thick granular sherds. Coiled vessels contained mixed charcoal and coarse sand, with both surfaces cordmarked.

        There were bones of 20 mammalian species, mostly deer, but also carnivores, which showed that people hunted large herbivores and small omnivores. Birds comprised 30% of total bone, while other bone included carp, grass, and blue fish. Judging from the total number and variety of hunted and domesticated animals, both the hunting economy and animal domestication grew. Plenty of shells of turtles, univalves and mollusks showed a significant fishing-gathering economy, but most important were paddy rice husks in cultural levels. Electronic microscopy concluded that husks found in the 1993 excavation were wild rice (Oryza rufipogon) with signs of early human manipulation, while the husks found in 1995 had traits of wild rice and O. indica & japonica, i.e., partly domesticated. Yuchan Cave stone tools were all percussed, with no ground stone. Ceramics were Early Neolithic and pre-Pengtoushan, with animal bone C14 dates of 8327 BC and 7911 BC.

        The nature of the Xianrendong economy in incipient agriculture was "slash and burn". Trees were felled, sun-dried, burned, their ashes used as fertilizer. Seeds were spread on or in the ground without tilling or weeding, and the ripened crop was harvested immediately. Stone choppers or axes were major farming tools in Xianrendong, Tiao Huan Tong and Yuchan Cave. Stone hoe-like tools, bone spades, and chisels were all used for chopping and digging. In the south, "slash and burn" plants were mainly asexually reproducing root tubers, but vegetables and legumes were also being grown as garden plants. While Xianrendong rice grain analysis inferred that agriculture was at the "select, nurture and domesticate" stage, the Yuchan Cave area could also be designated similarly with respect to paddy rice. It did not represent Early or Middle Neolithic village agriculture because paddy rice agricultural growth was Late Neolithic in the Nanling area. This was based on: (1) humid climate when natural foods such as edible wild vegetation, soft-shelled animals, and fishes were plentiful, which delayed agriculture, and (2) this limestone area resisted Early and Middle Neolithic settlement. Thus, it was impossible to develop "village agriculture".
         While agriculture appeared in the central Yangtze plains 8,500-10,000 years ago, gathering, fishing, and hunting were still more important. At Xianrendong, most bone was herbivore, small omnivore, bird and fish, plus turtle, univalve and mollusk shell, showing that hunting, gathering, and fishing were very important in daily village life, with agriculture secondary.


        The 7,500-8,500 year-old Middle Yangtze Pengtoushan culture represents the Neolithic "Incipient stage. Pengtoushan culture in the lower Li River plain was early Middle Neolithic, and included the Pengtoushan type site, Bashidang and Lijiakang (Li Family Hill; Fig. 1). Pengtoushan vessels were mainly mixed clay and charcoal round bottom vessels, including deep abdomenal jars, tube-shaped cauldrons, basins, troughs and plates, and some with sandy clay supports. Small flint tools and large ground stone vessels were also found (4).

        Most carbonized rice husks and grains were on and within Pengtoushan sherds. Palynology revealed little data other than grain size of >37.5 microns and edge thickness of 10-11 microns, like modern paddy rice (5). Rice husks and grains occurred not only in Lijiakang sherds dug in 1989, but also in 8,000 year-old Bashidang sherds. Likely, a pile of rice straw or husks was burnt or left to decay, as a mass of contemporaneous rice pollen was in the ash pit. In 1995, several hundred carbonized rice grains and big husks were in screened samples in riverbank cultural levels in west Bashidang. In an area several dozen m northwest and 4.5 m subsurface, extensive organic matter in this ancient black peat had rough rice, lotus, water chestnuts, peaches and almost 100 kinds of other plants. There were also bamboo and wood ware, stone carvings, and animal bone. Awned and awnless rice grain was preserved so well due to silt burial that it looked fresh (6). Many sherds were in this level.

        Large amounts of paddy rice at Pengtoushan, Bashidang, and Lijiakang indicate that 8,000 year-old lower Li River horticulture passed the selection and domestication stage and evolved to agriculture. Pengtoushan culture was already entering the early hoe and plough stage of stable village agriculture. At the Bashidang village site, the surrounding moat and wall (7) clearly showed that village paddy rice agriculture already existed 8,000 years ago at 29–30 degrees N. Lat. on the middle Yangtze.


        In early developmental Neolithic, Yangtze paddy rice agriculture occupied several areas: Kanshi lower level on the lower middle Yuan and Li Rivers, early Chengbeixi culture in Hubei, Majiabing culture at Taihu Lake, and Hemedu culture (Fig. 1) in the Yuxiao area south of Hangzhou Bay. These early middle Neolithic cultures were mainly agricultural at the Developmental hoe-plough agricultural stage.

        Kanshi lower level had round bottom, curved-legged vessels, plus deep abdomenal (some dual-eared or handled) and flat bottom jars, and other vessel types. Stone tools were ground stone axes, spades, chisels, dishes, and chipped flint (8). Remains of paddy rice in Hujia Compound (9), Li County, Hunan, and Farm Cemetery Fort site (10) at Quian Liang Lake in Yoyang City were identified as 7,300-7,700 year-old Kanshi lower culture, while remains of domesticated livestock at Hujia Compound were water buffalo, pig, sheep, and other animals.

        Most Chengbeixi ceramics contained red-brown or gray-brown sand and charcoal, their most developed being round bottom pots. However, curved-legged basins with stands and two-handled jars were also very common. Stone tools were quite crudely percussed, with a few partly ground. Most flaked tools were rectanguloid or trapezoid axes, chisels, etc. Chengbeixi paddy rice was in Beiqi at Yiducheng City (11), Hubei, and Zhi Cheng North and Liulingxi City in Jieguixian County (13). C14-dated bone in T6 (3) was 6,800±80 years old (13).

        Majiabing culture ceramics contained mainly red-brown sandy clay and included legged cauldrons, dishes, basins, and jars. Crude stone tools were axes, grindstones for knives and chisels, etc. Most cultivated rough and paddy rice grain were in levels 3 & 4 of Luojiajiao (Luo Family Corner), Tongxiang County, Zhejiang. Level 3 & 4 excavation in six directions showed carbonized paddy rice grain, most without embryo. While a third were japonica, they were also indica. Of the 567 level 4 pollen grains, 97% was cultivated rice (14). While paddy rice was the only pollen type and concentrated, other cultivated plants were gourds and reeds used in mats and ropes. Domestic animals were pig, water buffalo and dog (15), etc. Level 4 at Luojiajiao C14-dated 6,900-7,100 years ago.

        Early Hemedu cauldrons with stands, jars, basins, trays, dishes, large bowls, and ladles were mixed charcoal and clay. Major farm tools were animal shoulder-blade hoes, with stone tools used in other tasks. Living quarters were on posts supporting planks over a low-lying swamp. Level 4 piles of 70-80 cm thick rough rice, husks, straw and grass were at 10 different Hemedu loci in a 400 sq. m area. Large quantities of husks were in clay abdomenal vessels, with analysis of rough straw indicating japonica. Small bottle gourd, lotus seed, and tribulus crops occurred, while palynology showed legumes. Domesticated pig, dog, water buffalo and sheep also existed. Level 4 economy included hunting, fishing, and gathering, as seen in wild fruit remains, skeletons of 50 animal species, and 1000 bone tools. Bones included deer, rhinoceros, Asian elephant, bear, otter, monkey and alligator. C14 dates on Hemedu level 4 wood were 6725±140 to 6960±100 years ago (16).

        The growth of paddy rice agriculture in Kanxi lower level, Chengbeixi, Majiabing, and Hemedu showed the following traits: When the Incipient Stage was around the north side of Dongting Lake, the Developmental Stage had spread over the mid-lower Yangtze Valley with an improved rice that evolved from the previous single variety. Irrigation and drainage technology also grew, as seen in small paddy field ponds in Caozheshan site lower level at Huxian in Jiangsu province, and probably facilitated by tidal movement. As Caozheshan lower level was almost contemporaneous with early Majiabing culture, irrigation and drainage already existed. Livestock domestication matched agricultural growth, and included dog, pig, water buffalo and sheep. Table 1 illustrates the growth and distribution of paddy rice agriculture along the Yangtze Valley.


        On the mid-lower Yangtze River, selectively bred-domesticated paddy rice was mainly on the Huaiyu Hills west of the river, but it was also at Yuchan Cave in the north Lingnan Hills of Hunan. As Yuchan Cave was in a low subtropical moist area, its location and ecology made early rice agriculture impossible, while archaeological data showed that Middle Neolithic paddy rice agriculture was undeveloped. Early Middle Neolithic paddy rice was also absent in the area around Poyang Lake to Huaiyu Hills. However, 7,500-8,500 year-old Incipient stage paddy rice was in Pengtoushan culture on the northwest shore of Dongting Lake. Until now, cultivated rice of this period was absent on the lower Yangtze, but paddy rice in Majiabing and early Hemedu cultural levels were several dozen cm thick. Husks, straw and leaves were inside mixed charcoal and clay sherds. Paddy rice included indica & japonica, which showed that agriculture evolved from Incipient to Developmental Stages. Agriculture grew in late Hemedu and Majiabing cultures, and the following Zhongze and Liangzu cultures. From Majiabing to early Hemedu culture, a well-developed agro-culture had already begun on the lower Yangtze, but the "select-breeding-domesticate" stage was absent in its mid and east sections, as was incipient Pengtoushan culture on the middle Yangtze. Whether lower Yangtze paddy agriculture spread from the middle valley or grew locally needs further study and research. The search for the origin of lower Yangtze paddy rice should be coordinated with local Early Neolithic site survey, i.e., Early Neolithic sites must be found before Neolithic cultivated paddy rice sites. Neolithic research should be conducted on lower Yangtze Early Neolithic sites on the east and north hills of Tianmu and north hills of Ximing. West of the middle Yangtze, especially on the east Wuling Hills near Dongting Lake, there were Late Palaeolithic rice remains and Early Neolithic sites. Early paddy rice agriculture had already formed in 8,000 year-old Pengtoushan culture, but grew further in cultures of lower level Kanzi, Daxi, Jujialing and Shejiahewen. Whether this area represents the origin of Yangtze River paddy rice agriculture remains questionable for academics.


(1) Lui Shi Zhong. Important progress at Xianrendong and Diaotonghuan sites, Jiangxi province.

(2) Yuan Jia Rong. New proof of the origin of paddy agriculture in Yuchan Cave. Zhong Guo Wenwu, 8th issue, p. 1. Published 3rd March, 1996.

(3) Zhang Zhi Heng. Effects of ecology and environment on prehistoric cultures. Jiang Han Archaeology, 3rd issue, 1996.

(4) Hunan Institute of Archaeology: Report on Early Neolithic excavation at Pengtoushan (Lixian, Hunan). Zhong GuoWenwu, 8th issue, 1990.

(5) Palynology Laboratory, Hunan Institute of Archaeology: Pollen analysis in Pengtoushan site in Lixian, Hunan, and research on ancient environment, Wenwu, 8th issue 1990.

(6) He, Jie Jun. Discussion of Primitive Cultures of the middle Yangtze River, Collections of essays from the 2nd Asian Prehistoric Cultural Research Convention on Prehistoric Cultures of Middle Yangtze Valley, Yue Li Book Company, pp. 252-260, 1996.

(7) Wong Hong Yuen. Observations of early cities on the middle Yangtze River. Discussion of prehistoric cultural anthropology of the middle Yangtze River during the Second Session of Asian Civilization. Yue Li Book company, pp. 252-260, 1996.

(8) Hunan Provincial Museum, Hunan. Neolithic Archaeology in Kanxi of Ximunxian, Archaeology, 1st issue of 1986

(9) Hunan Institute of Archaeology. Neolithic site of Hu’ family compound in Lixian, Hunan. Archaeology Newspaper, 2nd issue of 1993.

(10) Yoyang City cultural feature work team: Attempted excavation of Neolithic site at Quian Liang Lake cemetery, Hunan Archaeology publications, 6 issues, Jia Quan He:Rediscussion of the theory of Middle Yangtze River primitive culture, Second session of Asian Civilization Symposium on Discussion of Middle Yangtze River prehistoric culture. Yue Li Book company, p. 205, 1996.

(11) Chen Zheng Yue: Cheng Bie Xi, Ye Du , Hubei. Prehistoric Research (Edited Publications), 1989.

(12) Hubei Institute of Archaeology: 1982 Early Neolithic Archaeological excavation in Luilin Xi county, Jiang Han Archaeology, 1st issue of 1994.

(13) Same as (11), p. 92

(14) Luo Family site Archaeological team. Report from the excavation at Luo’ Family Corner in Tong Xiang Xian, Zhejiang Province cultural feature archaeology studies publication, Cultural Publication House,1981.

(15) Zhang Ming Hua: Fauna at Luo’ Family Corner, Zejiang Provincial cultural Archaeology studies publication, Cultural Publication House, 1981.

(16) Zejiang Provincial Cultural Management committee:1st report on excavation at Hemedu, Archaeology Journal, 1st issue of 1978; Hemedu archaeology team: Main results at the 2nd excavation at Hemedu, Cultural features, 5th issue, 1980; Natural History Division of Zejiang Provincial Museum: Research on the appraisal of zoological and botanical remains at Hemedu. Archaeology Journal, 1st issue 1978.