Veget Hist Archaeobot (2000) 9 : 161-168

VegetationHistory and Archaeobotany 9 Springer-Verlag 2000

Pollen and phytolith evidence for rice cultivation during the Neolithic at Longqiuzhuang, eastern Jianghuai, China Fei Huang ~ and Min Zhang 2

i Nanjing Institute of Geology and Palaeontology, Chinese Academy of Sciences. Nanjing 210008, PR China 2 Institute of Archaeology Nanjing Museum, Nanjing 210016, PR China Received July 7, 1999 /Accepted January 10, 2000

Abstract. Phytolith and pollen analyses were carried out at the archaeological site at Longqiuzhuang in Gaoyou, Jiangsu, southern China. The results indicate that the key morphological phytolith types associated with cultivated rice (Oryza) are common in the Neolithic cultural layers at this site. The evidence strongly suggests that cultivated rice (mainly O. japonica) was grown locally during the Neolithic. The archaeopalynological record provides information about the impact of human activity and, in particular, farming on the natural vegetation. The evergreen and deciduous broad-leaved forest was substantially altered, and herbaceous taxa, including ruderals, expanded. Based on the results from the phytolith and pollen analyses, two distinct phases of human activity have been recognized, namely (1) phase A (7000-6300 B.P., i.e. early Neolithic) a warm and humid period when arable farming, including rice cultivation, was pursued but the variation in the size of the carbonized rice grains was low, and (2) phase B (6300-5500 B,P., late Neolithic age), a period of relatively cold and/or arid climate when cultivated rice was of major importance and was morphologically similar to present-day rice. Environment, and in particular climate change in the late Neolithic, were important factors affecting the development of rice as a cultivated crop. It was mainly during this period that artificial selection favoured the emergence of forms similar to those of today. Key words: Phytolith analysis - Pollen analysis - Rice

cultivation - Neolithic- Southern China

Introduction

During the course of the last few thousand years, human impact on natural vegetation complicates the interpretation of climatically induced changes as recorded in pollen diagrams. For several decades now, palynologists have taken human impact seriously into consideration when explaining Holocene vegetation history. Some authors re-

Correspondence to: F. Huang; e-mail: [email protected]

gard human activity as one of the main factors, or even the most important factor, determining vegetation dynamics during the Holocene. The human factor obviously complicates the application of pollen analysis to the elucidation of natural vegetation and climate changes. On the other hand, it is now fully accepted that pollen analysis is a powerful tool for the detection and quantification of the impact of past human activity. Pollen studies from archaeological sites, in particular, have provided much information on the interaction between human activity and vegetation. By comparing pollen spectra from archaeological sites, i.e. the local or disturbed situation, and nonarchaeologicM sites such a bogs and lakes, which reflect regional and natural and also human-influenced vegetation in a given area, it is possible to reconstruct in detail the transition from natural vegetation to a situation where farming activity is a major factor influencing vegetation dynamics. Even though traditional archaeology can furnish evidence for agricultural activity, phytolith analysis has brought new insights into environmental archaeology during the last decade. This includes providing an indirect but reliable method for detecting rice cultivation at archaeological sites (Rovner 1986; Piperno 1988; Rapp and Mulholland 1992; Pearsail and Piperno 1993; Fujiwara 1993). In the Jianghuai region, palynologists have paid much attention to the human role in Holocene vegetation dynamics, but the precision is not yet available to describe the course of vegetation changes that had a bearing on the early stages of rice cultivation. The Longqiuzhuang site, the results from which are described here, offered a rare opportunity to clarify human impact on vegetation and also to elucidate possible relationships between agricultural activity and environmental changes. At archaeological sites, fossil pollen is often poorly preserved depending on the local environmental context at a particular site. On the other hand, developments in phytolith analysis makes this a powerful tool in reconstructing past environments especially where conditions are not particularly favourable for pollen analysis. Phytoliths are generally abundant in prehistoric environments. Hence, phytolith research is not only a viable alternative to pollen studies in palaeoenvironmental recon-

162 struction but can contribute substantially on its own rights (Lewis 1981; Pearsall and Michael 1984; Piperno 1984, 1988) as is demonstrated in this particular instance.

Natural environment and archaeological context The Longqiuzhuang site lies in the eastern Jianghuai plain to the east o f Lake Gaoyou and the Jinghang canal (32~ 119~ 2.6 m asl; Fig. 1). The annual average temperature is 13.5-15.5~ January average temperature is in the range -1 to 2,5~ and the absolute minimum temperature is below -20~ Annual precipitation is in the range 900-1200 mm with most rain in the period June to September. The area lies within the northern subtropical evergreen and deciduous broad-leaved mixed forest (Wu,

1980). Because of human activity, the original vegetation is rarely to be seen except on the hills in the southern part of the study area, where there is northern subtropical mixed deciduous broad-leaved and evergreen forest. Characteristic tree taxa include deciduous species o f Quercus (e.g.Q. variabilis, Q. acutissima, Q. fabri), Cas-

tanea, Liquidambar, Dalbergia hupehana, Pistacia chinensis, etc., and broad-leaved evergreen taxa such as Cyclobalanopsis glauca, llex chinensis, Castanopsis, Ligustrum lueidum. Introduced trees associated with towns and villages include Salix matsudana, S. babylonica, Populus simonii, Ulmus pumila, Cetis sinensis, Photinia serrulata and Ligustrum lucidum. The main crops in the region are rice and wheat. The archaeological site, which is square in shape and surrounded by lake water, occupies some 50 000 m z. It is excellently preserved and is the largest Neolithic site in the Jianghuai area. The site was excavated four times between i 993-1995. In all, an area of 1335 m 2 was excavated in which eight distinct layers were recorded. Layers 4-8, generally 1 m in thickness, are o f Neolithic age. In trench GLT1524, six soil samples were collected from the cultural layers for pollen and phytolith investigations. Based on the cultural remains at the site and by comparison with corresponding cultural phases in the neighbouring areas, i.e. mainly in Taihu, Ningzhen and central Jianghuai, the following cultural phases are recognised at the Longqiuzhuang site: phase 1, i.e. layers 7-8 (7000-6300 B.P.) and phase 2, i.e. layers 4-6 (6300-5500 B.P.; Fig. 2).

N

[~0. 110

115

120

J

\-~

o

6

125CE

o

2o

s_s_%0p~ 6000 B.P.

a3020B.p. 7000 B,R

>7000B.P,

40kin 33

{

~L~

( i % Longqiuzhualg site

Fig. 2. Profile of the west wall in trench GLTI524, Longqiuzhuang, showing sampling points (indicated by open circles), layer numbers and age estimates

Laboratory methods Phytolith sample preparation 32

I18

120

Fig. 1. A Map of eastern China and adjoining areas. The main rivers, towns (open circles) and main sites where palaeoecological investigations have been carried out (closed circles) are indicated. 1, Longqiuzhuang; 2, Jiahu; 3, Yangzhuang; 4, Caoxieshan; 5, Longnan; 6, Hemudu; 7, Pengtoushan; O, Jianhu profile; | Qidong profile. B Detailed map showing the location of the Longqiuzhuang site, Jiangsu province

Samples of ca. 20 g wet weight were taken from the main archaeological layers (Table 1) for phytolith analysis. Samples were treated initially with 10% cold HC1 to remove carbonates, then with H202 to get rid of the organic matter and finally with Na2HCO 3 to defiocculate the humic acid. Each step in the above processes was followed by centrifugation and washing with water. After deflocculation, samples were subjected to ultrasonic treatment and repeatedly washed with water until neutral. Finally, the phytoliths were floated on a heavy liquid (specific gravity

163 Table 1. Main stratigraphical features at the location where samples were taken for pollen and phytolith analysis. Depths at which the samples LT-4 to LT-8 were taken are also given

Layer Stratigraphy No.

lowed by 37% HCI to dissolve siliceous rides and sieving with a sieve with a 5-]m treatment was ineffective in r e m o v i n g heavy liquid treatment (specific gravity out.

matter and fluomesh. Where HF mineral matter, 2.2) was carried

Results

1 2 4 5 6 7

8 9

Modern cultivation layer, greyish brown clay Yellow clay mixed with greyish clay Greyish black clay mixed with brownish yellow clay LT-4 taken at depth 0.55 m Greyish clay mixed with yellow sands LT-5 taken at depth 0.7 m Greyish black clay, locally rich in ashes, charcoal LT-6 taken at depth 0.9 m Yellow clay mixed with greyish sands and containing orignal shells at the bottom, secondary shells in the middle and upper part of the layer LT-7 taken at depth 1.15 m Greyish blown silt, containing reed rhizomes LT-8 taken at depth 1.4 m Yellowish grey silt, mixed with fine sands

2.30) and microscopically investigated at x400 magnification (Wang et al. 1993). Over 300 phytoliths were counted in each sample.

Pollen sample preparation Samples o f ca. 30 g wet weight were used for pollen analysis. Standard methods were used including treatment with 10% cold HC1 to remove calcium carbonate, HF fol-

Phytolith assemblages Abundant phytoliths were present and include the following forms: bilobate, fan (bulliforms), square/rectangular (bulliforms), smooth elongate, elongate, saddle, bamboo saddle, cross, trilobate, polylobate, pointed and tracheidshaped (Table 2; Fig. 3). The phytoliths were mainly o f grass origin. Epidermis cells were also recovered. Diatoms were present, which suggest a damp habitat around the site or, alternatively the diatoms may have been brought into the site from the nearby lake as part o f attempts at irrigation by the local inhabitants. The key phytoliths of Oryza, e.g. fan-shaped bulliform and bilobate-shaped phytoliths are c o m m o n in the cultural layers (Fig. 3, Nos. 3, 17). The bulliform-shaped phytolith is typically fan-shaped, with fish-scale-like ornamentation on the top and two lateral protrusions. These are regarded as the criteria for identification o f phytoliths associated with cultivated rice in Japan (Fujiwara 1993). The bilobate, i.e. dumbbell-shaped, phytoliths, with concave ends and the long axis o f the indivudual phytoliths running parallel to each other and perpendicular to the leaf vein, are characteristic o f the tribe Oryza (Kealhofer and Piperno 1994; Jiang 1995).

Table 2. Phytolith statistics of samples from trench GLT1524, Longqiuzhuang site

Early Neolithic Sampling number Frequency

Phytolith type Bilobate* Normal bulliform Oryza-type bulliform Reed-type bulliform Square/rectangular Smooth elongate Elongate Saddle Bamboo saddle Cross Trilobate Point Tracheid Sedge-type phytolith Others Epidermis fragments Diatoms Sum

LT-8

LT-7

LT-6

AF

RF

AF

RF

AF

RF

88 28 1 2 21 26 21 60 40 2 5 18

27.0

124 60 3 1 34 32 33 34 24 3 12 19

28.1

136 3 1 1 10 82 14 10 35 6 13

36.8

10 3 1 326

9.5 6.4 8.0 6.4 18.4 12.3 0.6 1.5 5.5 3.1 1.0 0.3

7 6 50 442

1.4 7.7 7.2 7.5 7.7 5.4 0.7 2.7 4.3 1.6 1.4 11.3

1.4

1

2.7 22.2 3.8 2.7 9.5 1.6 3.5 0.3

5 11 42

1.4 2.9 11.4

370

Late Neolithic LT-5 AF RF

LT-4 AF

RF

177 5 1 1 21 78 23 8 17 2 6

47.3

126 10 3

38.7

23 60 29 3 19 2 7

7.1 18.4 8.9 0.9 5.8 0.6 2.1

12

3.7

7

1.9

1 1 7 12 11

0.3 0.3 2.1 3.7 3.4

3 3 2 17 3

0.8 0.8 0.5 4.5 0.8

3.9

326

*Bilobate and characteristic of the tribe Oryza; AF: absolute frequency; RF: relative frequency, i.e. % of total

374

1.9 5.6 20.9 6.1 2.1 4.5 0.5 1.6

164

Fig. 3. Phytoliths types recorded at Longqiuzhuang site. 1-4, 12, 28-30, Fan-shaped bulliforms (3, Oryza-type bulliform); 5, 6, I3, point-shaped; 7, elongate; 8, polylobate; 9, trilobate; 10, smooth elongate; 1I, 14, tall saddle-shaped; /5, 17, /8-20, 23, 24, 26, 27, bilobate (17, bilobate and characteristic of the tribe Oryza; 16, 31, square-rectangular bulliforms; 21, cross-shaped; 22, saddle-shaped; 25, bamboo saddle-shaped

165 The presence of phytoliths of Oryza is consistent with the archaeobotanical evidence for rice cultivation during the Neolithic. The mean ratio of the length of the lower part (a) to the upper part (b) in Oryza-type buIliforms is <1 (Fig. 4), which is the case in the material recorded here. This indicates that we are dealing with cultivated rice, i.e.O, japonica (Fujiwara 1978).

Fig. 4. Main morphologicla parameters of fan-shaped bulliform The frequent occurrence of carbonized rice grains with artefacts of Neolithic age in China indicates that China is one of places of origin of cultivated rice in Asia. In the 1970s, at the Hemudu site in Zhejiang, carbonized grains and stone tools used in rice cultivation were recovered in the course of an excavation (Natural History Section, Chekiang Provincial Museum 1978) and many other Neolithic sites with evidence for rice cultivation were discovered in the middle-lower reaches of the Yangtze river (You 1986). There are several centres which undoubtedly contributed at an early stage to the genetic diversity of Chinese cultivated rice. These include centres in the middle and lower reaches of the Yangtze river and southern China including Yunnan province (Yan 1989; Di 1957; Liu 1975). Discoveries of other sites with rice cultivation dating to the Neolithic have been made in recent years, e.g. Pengtoushan in Human (Pei 1989), Jiahu in Henan (Cultural Relic Institute of Henan province 1989) and Longqiuzhuang in Jiangsu. These discoveries expand greatly the known region of rice cultivation to include the mid-Yangtze valley and the upper Huai river regions. The Jiahu site in particular, which has been 14C dated to 7900 cal. B.P., suggests that rice cultivation in China began 1000 years prior to the Hemudu period which dates to ca. 7000 cal. B.P. (6950+150 cal. B.P.), i.e. the early Neolithic (Chert et al. 1995; Kong et al. 1996). It is assumed that the common wild rice (Oryza rufipogon) is the ancestor of cultivated rice (O. sativa, including O. indica and O. japonica). The morphological differences between O. indica and O. japonica are small so that these are distinguished only with difficulty on the basis of macroscopic features. Phytolith analysis can have a useful role here, however, in that the phytoliths in O. indica and O. japonica can be easily distinguished on morphological criteria (details in Sato 1990). In China, phytoliths have been widely applied in archaeological research on rice cultivation during the Neolithic in recent years. Phytolith analysis at the Longnan site in Jiangsu indicated that the cultivated rice

was mainly O.japonica (Tang 1992). At Hemudu, the rice remains which showed great variation in the spikelets suggest a mixed population (Zhen et al. 1994). At Jiahu, O. japonica was the major component and there was some O. indica (Chert et al. 1995). The abundance of typical rice phytoliths at the Yangzhuang site in Henan suggests rice cultivation (Jiang 1995). At Caoxieshan in Jiangsu, mainly O. japonica was recorded (Huang et al. 1998). Based on the phytolith data now available from carbonized rice grains, it appears that O. japonica records relate mainly to the middle reaches of the Yangtze river and the upper reaches of the Huai river regions. During the course of excavation at Longqiuzhuang, over 4000 carbonized rice grains were recovered by the wet sieving method from the Neolithic cultural layers 8-6 and 4 (Zhang 1997). The results of grain size measurements are summarized in Table 3. As rice cultivation developed, size variation of carbonized grains increased. In the early Neolithic period (7000-6300 B.P.), there was minimum variation and so it can be concluded that the evolution of cultivated varieties occurred slowly. In the late Neolithic (6300-5500 B.P.), maximum variation is recorded which indicates that effective artificial selection had taken place. Phytolith analyses of soil samples from different environments provide data relevant to the reconstruction of local palaeoenvironmental conditions. According to Wang et al. (1993), the bulliform and bamboo saddleshaped phytoliths are mainly distributed in south China (subtropical-tropical climate zone) while bilobate-shaped phytoliths occur mainly in the Jianghuai area (northern subtropical climate zone) and elongate-shaped (including smooth elongate) phytoliths are mainly associated with plants adapted to cold environments. Bilobate and bulliform-shaped phytoliths, which mainly occur in warm and humid environments, predominated in the earlier phases at Longqiuzhuang but a substantial change occurred between the early and late Neolithic (Fig. 5). In recognition of this, two climatic Table 3. Mean, standard deviation and coefficient of variation of carbonized rice grains from Longqiuzhuang (after Zhang 1997)

Archaeological layer No. of carbonized rice grains

-~

L-8 14

Mean (ram) 4.48 Standard deviation 0.47 Coefficient of variation (%) 9.65

L-7 65

L-6 48

L-4 118

4 . 7 2 4.58 5.8 0 . 5 6 0.51 0.69 11.9 11.13 11.87

Mean (mm) 2.24 2 . 3 2 2 . 2 8 2.57 Standard deviation 0.23 0.31 0.3 0,45 Coefficient of variation (%) 10.1 13.61 13.09 17.86 ~ _~

Mean(mm) 1.65 1.69 1.65 1,78 Standarddeviation 0.21 0 . 2 3 0 . 2 9 0.41 Coefficient of variation (%) 12.74 13.67 17.64 23.03

166

Pollen and spores

" ~" ~ ~" ......... .r ............... . . . .

......... :"I ......... . . . . . I~" ..

4.~:--"--,~"'~"-~

; ' - - I

I

!

I

......

F ""

!

--

':" ""~ =' ~

z.

Phytoliths and other microfossils

. . . . .6 . .1. . . . . . . . . . .

t-

i~

H

9

9

' L....................

ix-.'.:_'_~l..~!

F"

!

i 6, ............... ~............ i m

a i

..... i . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

-* ........

' l

l I

I ~ " J

........ t...~ I

:

l

I

i

i I

....~ .... """.."z. i

I

I............. , I

9

Ph~soB1

I

....... l ........ m ........ L . . I .................. I... L . ~ . . . . I

'

.............

i

!

1

/

7' I

I

I

I /

[

I i

8- t ~ . ~ - t ~

-

~

, ~

I

I i

.' ~a,- R44 }+~-.

I

I

I

:1

I

l

!l '1

l I1

I.

l

I

L

I

!

I Pha~eA]

:

~ - ~ . ~ 20

40

2..~;~, ~ - ~ -

~ ,~-

~

20

~ . ~- ~ . ~. ~ e~.

~.~W~-v 20 40

2~

~

~

20

. ~" m , ~ " ~ , ~ . 20 20

~" "~" ~,. ~ " ~ - ~ . ~ ' ~ ' l

20

............

Fig. 5. Diagram showing percentage representation of the main pollen and phytolith taxa recorded at the Longqiuzhuang site. In the pollen diagram a total pollen sum is used. The other curves (mainly phytoliths) are expressed in terms of the sum of all taxa in this grouping phases are distinguished with human impact and agricultural activity. In phase A (layers 7-8) in which bilobate(25.7%), bulliform- (including normal, Oryza-type, reedtype; 14.5%) and bamboo saddle-shaped (13.5%) phytoliths dominate, which suggests a warm and humid climate during which rice cultivation was pursued. The high representation of bamboo saddle-shaped phytoliths suggests that bamboo forest was used to a considerable extent in this period. Phase B (layers 4-6) is characterized by a substantial decrease in bulliform and bamboo saddleshaped phytoliths and predominance of bilobate- (mainly from Panicoideae; 40.9%), elongate- (including smooth elongate; 25.6%) shaped phytoliths. This suggests a relatively cool and/or arid environment, and the beginning of artificial selection because of environment pressure. Archaeopalynological record

Holocene vegetation development is known from archaeopalynogical and stratigraphic data available from the Jianghuai area (Zhang 1982, 1983; Tang et al. 1983; 1992; 1996; Liu et al. 1992). There is a high resolution Holocene pollen profile from Jianhu, 80 km north-east of Longqiuzhuang (Tang et al. 1992, 1996) and the Qidong profile from 230 km to the south-east is also important (Liu et al. 1992). However, the relationship between human activity and vegetation development has still to be fully clarified. In this context, the Longqiuzhuang site is particularly important. According to Behre' (1981, 1990), primary anthropogenic pollen indicators derive from cultivated plants and particularly cereals. Secondary anthropogenic indicator includes non-arboreal pollen (NAP) and especially pollen of ruderals and weeds associated with arable activity, i.e. plants not intentionally favoured by humans. Thus, with the onset of farming and the opening-up of the forests, plants including Plantago, Polygonum, Urtica, A r t e m i s i a and Chenopodiaceae expand and register often for the first time in Holocene pollen diagrams. At Longqiuzhuang, in samples LT-7 and LT-8, i.e. the early Neolithic cultural layers, pollen was plentiful. AP accounted for

only 17% so that the spectra consist predominantly of NAP (78.7%; Table 4, Fig. 5). AP consists mainly by C y c l o b a l a n o p s i s and Q u e r c u s (<5%), while Pinus, Ulmus, Liquidambar, Betula, C o r y l u s + C a r p i n u s (combined) and J u g l a n s together do not exceed 10%. Pollen of other trees such as Castanopsis, Tilia, Taxodiaceae and Cupressaceae are only sporadically recorded. Among the herbaceous taxa, Gramineae (44%), A r t e m i s i a (18.7%) and P o l y g o n u m (10.3%) predominate, while Chenopodiaceae, Ranunculaceae and aquatic taxa, e.g. Typha and Sparganiaceae, are less important. Pteridophyta and bryophyta are presented only by single spores. In samples LT-6, LT-5 and LT-4, i.e. late Neolithic cultural layers, pollen concentration is low and so the pollen sum is less than 200. In these samples, however, Oryza pollen, which has a diameter >40 im (Sun et al. 1981 ; Wang et al. 1996), is often present. In sample LT-4, NAP is dominated by Gramineae, Compositae and Chenopodiaceae. In summary then, the archaeopalynological record indicates that damp conditions prevailed at the Longqiuzhuang site and that only fragments of forest were present as the cultural layers, which are referable to the Neolithic, accumulated. The Jianhu pollen profile from 80 km north-east of Longqiuzhuang (Tang et al. 1992; 1996) shows a regional vegetation characterized by evergreen and deciduous broad-leaved forest ( C y c l o b a l a n o p s i s , Castanopsis, Quereus, Liquidambar, Ulmus and Pinus) between 85003700 B.P. which corresponds broadly with what is regarded as the Holocene thermal maximum in China (Shi 1992). The situation is similar in the Qidong pollen profile (Liu et al. 1992). However, from ca. 6400 B.P. (possibly equivalent to the late Neolithic) conifers expanded. This phenomenon has also been documented in the Taihu area. On the basis of three pollen profiles from Taihu lake, Xu et al. (1996) showed that evergreen broad-leaved trees declined and conifers, together with warm-temperate elements, increased from the late Neolithic (about 60005000 B.P.) onwards. This is probably the result of human impact on the vegetation. Liu (1988) and Liu et al. (1992) suggest that both climatic cooling and human disturbance may have contributed to the late Holocene Pinus rise de-

167 Table 4. Pollen and spores statistics of samples from trench GLT1524 at the Longqiuzhuang site

Sampling no. Frequency

Early Neolithic LT-8 LT-7 AF RF AF RF

Late Neolithic LT-6 LT-5 LT-4 AF AF AF

Taxon names Pinus Quercus Cyetobatanopsis Betula Corylus+Carpinus Ulmus Liquidambar Juglans Alnus Fagus

Taxodiaceae Rhamnus

Arboreal pollen Gramineae* Artemisia

9 3 6 4

3.3 1.1 2.2 1.5

4 6 9 3

1.6 2.4 3.6 1.2

3

1.1

9

3.3

3

1.2

2 2

O.7 0.7

10 2

4 O.8

3 2

1. l 0.7

2

0.8

5 1.9 1 0.4 49 18.2 39 15.7 114 42.4 114 45.8 55 20.4 42 16.9

Compositae Polygonum

Chenopodiaceae Ranunculaceae Labiatae Rubiaceae

26 7 1

9.7 2.6 0.4

Humulus

Liliaceae Typha

Sparganiaceae Potamogetonaceae NAP Botrychium Ceratopteris Microlepia

Fern spores Bryophytes spore Total

207 7 10 3 269

0.4 0.7 0.4

4 13 2 1

2

2

1

1

7 26 9 1 2 3

-

3 1

1.2 0.4

3 17 52 5 10 1 6 -

-

O.8

77.0 200 80.3 2.6 3.7 1.1

2 2

Conclusions

27 10.8 7 2.8 2 0.8 1 0.4 1 0.4 2

1 2 1

4

1 2

2

2 18

43

1 2

6 86 3

3

1.2

-

2

I 4 6

0.4 1.6 2.4

1 1

2

3 12

23

52

111

249

reaches of the Yangtze river; 3-4~ in northern China; and 4-5~ in Tibet. Annual precipitation was also different. In the south-east monsoon area, for instance, there was at least 100 mm more precipitation than at present. Under these favourable climatic conditions, human culture flourished, agriculture developed and human disturbance of the natural ecosystem increased gradually. The archaeopalynological results from Longqiuzhuang show that the natural plant cover was strongly affected by human activity. Forest regeneration after human disturbance was marked by trees indicative of forest succession, e.g. pine. Early rice cultivation was the main cause of the vegetation changes observed in the Holocene pollen records from the Jianghuai region. The evergreen and deciduous broad-leaved forests were substantially altered and ruderals spread in the deforested areas during the late Neolithic.

Additional taxa (count given in each instance) LT-4: Castanea + Castanopsis 3; Tilia 1; Cupressaceae 3; Umbe[liferae 1; Caryophyllaceae 1; Thalictrum 1; LT-5: Oteaceae 1 * Oryza pollen observed; AF, numbers counted; RF, percentage representation tected in the pollen record from northern China and northeast China. During the thermal maximum (ca. 8.500-3000 B.P.; see above), the vegetation zone boundaries shifted northwards so that the steppe boundary lay to the north o f its present position, the warm-temperate broad-leaved forests shifted northwards to 40-50~ in eastern China and the subtropical evergreen broad-leaved forest lay at ca. 35~ in the Qinling mountains (An et al. 1991; Sun and Chen 1991). Lake levels and sea levels may also have risen during this period. During the Holocene climatic optimum, temperature were probably higher than at present as follows: I~ in our study region; 2~ in the lower

The development o f farming based on rice cultivation at Longqiuzhuang may be divided into two phases. In phase A (early Neolithic, 7000-6300 B.P.), the climate was warm and humid and the already partially open landscape with wet habitats was used for rice growing. The small size variation in the carbonized rice grains suggest that the rice cultivation was primitive. In phase B (late Neolithic, 6300-5500 B.P.), there is evidence for artificial selective breeding which took place in the context of changing environmental conditions. Variation in grain size reached its maximum, the sizes are similar to those of today and the vegetation around the site was intensively altered by human activity, i.e. the evergreen and deciduous broadleaved forest was reduced in extent to give a largely open landscape. Acknowledgements. This study was supported by the Labora-

tory of Palaeobiology and Stratigraphy, Nanjing Institute of Geology and Palaeontology, Chinese Academy of Sciences (Grant No. 980105). The authors are indebted to Prof. S. Ouyang for critically reading the manuscript and for helpful suggestions. We thank Dr. W.M. Wang, Prof. F.B. Wang, Prof. H.Y, Huiyou and Prof. J.L. Liu for interesting discussions, and Mrs. F.B. Huang for pollen and phytolith extraction.

References

An ZS, Wu XH, Lu YC (1991) Paleoenvironmental changes of China during the last 18000 years. In: Liu TS (ed) Quaternary geology and environment in China. Science Press, Beijing, pp 228-237 Behre KE ( 1981) The interpretation ofanthropogenic indicators in pollen diagrams. Pollen Spores 23:225-245 Behre KE (1990) Some reflections on anthropogenic indicators and the record of prehistoric occupation phases in pollen diagrams from the Near East. In: Bottema S, Entjes-Nieborg G, Zeist W van (eds) Man's role in the shaping of the eastern Mediterranean landscape. Balkema, Rotterdam, pp 219-230 Chen BZ, Wang XK~Zhang JH (1995) The discovery and significance of Oryza phytoliths in the Neolithic site at Jiahu, Ouyang County, Henan Province (in Chinese). Chinese Sci But! 40: 339-342

168 Cultural Relic Institute of Henan Province (1989) Reports on the 2"<6 th excavations of Neolithic site in Ouyang county, Henan province (in Chinese). Cult Relics 1: 1-4 Di Y (1957) The origin and development of rice cultivation in China (in Chinese). J Agric (Beijing) 8:243-260 Fujiwara H (1978) Fundamental studies in plant opal analysis (2): the shape of the silica bodies of Oryza (in Japanese). Archaeol Nat Sci (Tokyo) 11 : 9-19 Fuj iwara H (1993) Research into the history of rice cultivation using plant opal analysis, ln: Pearsall DM, Piperno DR (eds) Current research in phytolith analysis: applications in archaeology and paleoecology. MASCA Research Papers in Science and Archaeology, Vol. 10, University of Pennsylvania Museum, Philadelphia, pp 147-158 Huang F, Wang WM, Li MC (1998) Phytolith analysis at the Neolithic site at Caoxieshan, Suzhou (in Chinese with English abstract). Acta Micropalaeontol Sinica 15:78-85 Jiang QH (1995) Searching for evidence &early rice agriculture at prehistoric sites in Chinathrough phytolith analysis: an example from Central China. Rev Palaeobot Palynol 89:481-185 Kealhofer L, Piperno DR (1994) Early agriculture in southeast Asia: phytotith evidence from the Bang Pakong Valley, Thailand. Antiquity 68:564-572 Kong ZC, Liu C J, Zhang QZ (1996) The discovery and significance of rice remains at a Neolithic site at Jiahu, Ouyang county, Henan province (in Chinese). Archaeology (Beijing) 12:78-83 Lewis RO (1981) Use of opal phytoliths in paleoenvironmental reconstruction. J Ethnobiol 1: 175-181 Liu KB (1988) Quaternary history of the temperate forests of China. Quat Sci Rev 7:1-20 Liu KB, Sun SC, Jiang XH (1992) Environmental change in the Yangtze river delta since 12,000 years B.P. Quat Res 38:32-45 Liu ZM (1975) The origin and development &rice cultivation in China (in Chinese). J Gen (Beijing) 2:21-29 Natural History Section, Chekiang Provincial Museum (1978) A study of the animal an d plant remains unearthed at Hemudu (in Chinese with English abstract). Acta Arch Sinica 1: 95-107 Pearsall DM (1978) Phytolith analysis of archaeological soils: evidence for maize cultivation in formative Ecuador. Science 199:177-178 Pearsall DM, Michael K(1984) Identifying past agricultural activity through soil phytolith analysis: a case study from the Hawaiian islands. J Archaeol Sci 11:119-131 Pearsall DM, Piperno DR (1993) Current research in phytolith analysis: applications in archaeology and paleoecology. MASCA Research Papers in Science and Archaeology, Vol. 10, University of Pennsylvania Museum, Philadelphia Pei AP (1989) Pengtoushan rice remains and prehistoric rice cultivation in China (in Chinese). Agric Archaeol (Jiangxi) 2: 102-108 Piperno DR (1984) A comparison and differentiation of phytoliths from maize and wild grasses: use of morphological criteria. Am Antiq 49:361-383 Piperno DR (1988) Phytolith analysis: an archaeological and geological perspective, Academic Press, San Diego Rapp GR Jr, Mulholland SC (1992) Phytolith systematics: emerging issues. Plenum, New York Rovner I (1986) Plant opal phytolith analysis in archaeology and paleoecology. In: Proceedings of 1984 phytolith research workshop. North Carolina State University, Raleigh. Occasional Paper No. 1 of the Phytolitharien: 1-147

Sato YI, Fuj iwara H, Udatsu T (1990) Morphological differences in silica body derived from motor cells of Oryza indica and japonica in rice (in Japanese). Japan J Breed 40:495-504 Shi SH (1992) The environment of the Holocene hypsithermal and Neolithic cultural development (in Chinese). In: Shi YF, Kong ZC, Wang SM, Tang LY, Wang FB, Yao TD, Zhao XT, Shi SH (eds) The climates and environments of the Holocene hypsithermal in China (in Chinese). China Ocean Press, Beijing, pp 885-181 Shi YF (1992) The climates and environments of the HHolocene hypsithermat in China (in Chinese). China Ocean Press, Beijing Sun X J, Du NQ, Chen MH (1981 ) The vegetation and climate in the Humudu period (in Chinese with English abstract). Acta Bot Sinica 23:146-151 Sun X J, Chen YS (1991) Holocene palynological research in China (in Chinese). In: Liu TS (ed) Quaternary geology and environment in China. Science Press, Beijing, pp 214-227 Tang LH (1992) Studies on rice cultivation at Longnan, Jiangsu (in Chinese). Agric Archaeol (Jiangxi) l: 70-73 Tang LY, Shen CM, Li MC (1983) The relationship between Neolithic culture and environment in northern Jianghuai area (in Chinese). In: Chou KS, Gong QM (eds) Researches in environmental archaeology 1 (in Chinese). Science Press, Beijing, pp 164-173 Tang LY, Shen CM (1992) Hypsithermal vegetation and climate in northern Jiangsu. In: Shi YF, Kong ZC, Wang SM, Tang LY, Wang FB, Yao TD, Zhao XT, Shi SH (eds) The climates and environments of the Holocene hypsithermal in China. China Ocean Press, Beijing, pp 80-94 Tang LY, Shen CM, Yu G (1996) Holocene vegetation history in the middle and lower reaches of Yangtze river (in Chinese). In: Shi YF, Zhang P1, Kong ZC, Gong YF, Go QH (ecls) Historical climatic change in China (in Chinese). Shangdong Scientific and Technological Press, Shangdong, pp 60-80 Wang KF, Zhang YL, Feng WQ (1996) Investigations into Holocene vegetation, environment and human activity in the Shanghai region (in Chinese with English abstract). Marine Geol Quat Geol (Beijing) 16:1-6 Wang YJ, Lu HY (1993) PhYtolith study and its application (in Chinese). China Ocean Press, Beijing Wu ZY (ed.) (1980) The vegetation of China (in Chinese). Science Press, Beijing Xu XM, Chang WYB, Liu JL (!996) Changes in vegetation and climate in the Taihu lake basin during the last 11 000 years (in Chinese with English abstract). Acta Palaeontol Sinica 35: 155-185 Yah WM (1989) The origin of rice cultivation in China (in Chinese). Agric Archaeol (Jiangxi) 1:19-31; 2:50-54 You XL (1986) The origin, dispersion and development of rice cultivation in the Taihu basin (in Chinese). China Agric Hist 1: 71-83 Zhang JE (1982) Quaternary pollen assemblage in Shenxian cave, Lishui, Jiangsu (in Chinese). In: Proceedings of the 1St China Palynology Congress, Science Press, Beijing, pp 32-38 Zhang JE (1983) Pollen analysis at Qingdun in Haian county, Jiangsu province (in Chinese). Acta ArchaeoI Sinica 2: 188-189 Zhang M (1997) The primitive rice cultivation in the Jianghuai plain (in Chinese). Agric Archaeol (Jiangxi) 1:106-120 Zhen YF, You XL (1994) Oryza phytolith analysis at the Hemudu site (in Chinese with English abstract). J Zhejiang Agric Uni 20: 81-85