1 Introduction

1.1 Background To the Survey

ERM-Hong Kong, Limited (ERM) has been commissioned by the Castle Peak Power Company Limited (CAPCO) to undertake the EIA for a proposed LNG terminal at South Soko Island and the associated cabling landing point at Shek Pik. A Cultural Heritage Impact Assessment (CHIA) is one of the components of the EIA study. To obtain field data for the CHIA, an Archaeological Survey at South Soko and Shek Pik has been undertaken.

The purpose of the archaeological survey is to investigate the presence of any archaeological deposits within 100 m from the boundary of the proposed development in Tai A Chau (see Figure 1.1) and the proposed cable and watermain landing point at Shek Pik (see Figure 1.2). Where archaeological remains are identified, their nature, horizontal and vertical extent has been determined.

Prior to archaeological survey commencement, a Licence to conduct the archaeological survey and an Excavation Permit were obtained from the Antiquities and Monuments Office (AMO) under the Antiquities and Monuments Ordinance (Cap. 53) and the District Land Office (Islands) under the Land (Miscellaneous Provisions) Ordinance (Cap. 28) respectively.

As the development boundary has been revised to include a larger area during the course of the Study as shown in Figure 1.1, two archaeological surveys was undertaken. The fieldworks were undertaken at South Soko Island between 5 September 2005 and 28 October 2005 and between 9 January and 17 January 2006. A previous preliminary survey was conducted for this Project in August 2004 where 4 test pits and 20 auger holes were excavated. The findings are summarised in Part 2, Section 12.3. This Report presents the findings for the archaeological surveys.

1.2 Archaeological Survey Team Members

The following individuals that participated in the archaeological surveys:

Mr Steven Ng of ERM Project Field Director, Licensed Archaeologist;

Ms Peggy Wong of ERM Trained Field Archaeologist;

Mr Wong Fu Trained Field Archaeologist; and

Ms Zoe Chan of ERM Trained Assistant Field Archaeologist.

In addition to the above team members, eight trained labourers were employed to assist in the surveys and a team of qualified Land Surveyors from Land Marker (1980) HK Company Limited assisted with the land surveying work.

Mr Steven Ng, Ms Peggy Wong, Ms Zoe Chan and Mr Wong Fu undertook post-excavation processing and analysis of fieldwork records, artefact assessment, preparation of rubbings, drawings, video recording and photographic records of artefacts.

1.3 Structure of the Report

Following this introductory section, the remainder of this report comprises the following sections:

Section 2 describes the objectives and methodology for the archaeological survey;

Section 3 presents the site background (including information on the geology, topography, hydrology, historical, ethnological and archaeological background);

Section 4 presents the archaeological survey findings;

Section 5 presents the preliminary finds assessment; and

Section 6 presents the conclusions.

The following annexes have also been included:

Annex A Soil Profile of Auger Holes;

Annex B Stratigraphy of Test Pits & Archaeological Features Records;

Annex C Soil Profile of Auger Holes and Stratigraphy of Test Pits for 2004 Survey; and

Annex D Land Survey of Test Pits Positions for 2005 and 2006 Surveys.

2 Objectives and Methodology

2.1 Objectives

The objectives of the archaeological surveys were to obtain adequate data to determine the presence, extent, depth, chronology, character and survival condition of identified archaeological deposits, if any.

2.2 Methodology

2.2.1 Desktop Study

Prior to fieldwork commencement, a desktop review and document-based research were undertaken to establish the fieldwork scope and strategy to obtain adequate field data, following Sections 1.4.3 and 1.5.3 of Criteria for Cultural Heritage Impact Assessment established by AMO. This desktop study included a review of site conditions, past land use, vegetation cover, landform, hydrogeology, fresh water sources, food and mineral sources, annual wind direction and ocean current patterns, previous archaeological fieldwork findings, historical and geological maps as well as 1950s to 1990s aerial photographs. In addition to the desktop review, site inspections were undertaken to evaluate the archaeological potential of the Study Area. Based on these findings, the fieldwork scope and strategy were agreed with AMO of Leisure and Cultural Services Department (LCSD) prior to fieldwork commencement is presented below.

2.2.2 Fieldwork Scope and Strategy

The fieldworks were undertaken in three main phases:

1. August 2004 Survey at Detention Centre at South Soko;

2. September to October 2005 Survey at proposed development area; and

3. January 2006 additional Survey arising from modifications to the project layout design.

The coverage of the survey areas is presented in Figure 2.1 and Figure 1.2.

Based on the desktop findings, the archaeological potential within the development boundary was evaluated identify the archaeological potential and decide the subsequent fieldwork strategy. The development area had been divided into seven sites, (Sites A to E and G at South Soko) and Site F at Shek Pik. The development layout was reviewed and the fieldwork focused on areas where potential soil disturbance may occur.

In evaluating the archaeological potential, a number of factors were reviewed including original landform (such as orientation, inclination of slope, elevation), food, freshwater and mineral resources availability, geology, sea currents, wind pattern, and vegetation to identify potential areas that favoured human inhabitation. Past land uses, natural erosion and previous archaeological investigations were also evaluated to establish the archaeological potential of the areas.

2.2.3 Field Walking

Field walking for surface collection was then conducted within the Project Area to assess the archaeological potential based on the distribution density of artefacts on the ground surface. The position of any special artefacts discovered was recorded. For special and general artefacts, their extent, quantity and chronology were recorded. The fieldwalk findings also assisted in refining the proposed test pit and auger hole locations.

2.2.4 Hand Augering

Augering was undertaken to investigate the vertical soil profile according to the soil colour or texture so that the presence of any cultural layers can be identified. Comparison of the stratigraphy of auger holes assists in determining the extent of an archaeological deposit area. A total of 56 (see Table 2.1) and 19 auger holes were bored respectively at South Soko and Shek Pik (see Section 4 for details).

2.2.5 Test Pits Excavation

The purpose of test pitting is to investigate the vertical and horizontal extent of any identified cultural layers. Three phases of test pit excavation were undertaken. The test pits undertaken are presented in Table 2.1. On South Soko island, the test pits were allocated at 6 sub sites (namely Sites A to E and Site G) and a total of 61 test pits were excavated (see Table 2.1). At Shek Pik, 7 test pits were allocated at the Shek Pik Tung Wan Archaeological Site and in proximity to the declared monument (Shek Pik Rock Carving) (see Section 4 for details).

Table 2.1 Quantity of Archaeological Fieldwork Undertaken between 2004 and 2006

Area	Test Pit	Hand Auger Hole
A	12	5
B	16	5
C	10	20
D	3	6
E	17	15
F	7	19
G	3	4
Total	68	74

The size of the test pits measured in the range of 1 m x 1.5 m, 2 m x 1 m, 2.2 m x 2.2 m, 2 m x 2 m, 1 m x 2.7 m and 3 m x 1.5 m subject to site condition. The test pits were excavated to depths ranging from 0.8 m to 3.5 m below ground. All test pits were excavated to the sterile layer. Most test pits were excavated manually by trained labourers except those in Site C on South Soko where a power tool hammer was used for concrete removal. During the excavation, trained labourers were employed under the supervision of the project field director or archaeologists. After completion of excavation works, all test pits were backfilled and reinstated to their original status.

2.2.6 Recording, Reporting and Processing of Finds

The site code of the archaeological surveys was designated as TAC05/TAC06 and SP05; TAC representing ‘Tai A Chau’(South Soko) and SP representing ‘Shek Pik’ respectively, and the year of the survey.

Test pit locations were recorded according to the Hong Kong metric grid system. A team of qualified land surveyors established the test pit locations. The site benchmark was tied to Hong Kong Principal Datum (mPD).

The stratum of each test pit or auger hole was distinguished by natural deposits in terms of soil colour, soil texture and any human activities or cultural remains. All auger holes, test pits and deposits were recorded using ERM’s recording system, which is compliant with AMO’s standard.

Sections drawings of the soil stratigraphy, and photographic records of at least one section of each test pit, were undertaken whenever site conditions allowed.

The formation of soil layers is influenced by both natural and human factors. A soil layer without human or artificial remains is generally classified as a “natural layer”. A soil layer with man-made features or remains is regarded as a “cultural layer”.

All artefacts identified were carefully washed, cleaned, labelled, bagged and boxed. Their functions and chronology were preliminarily assessed, and when possible, sorted typologically. Special or datable finds were registered, drawn and photographed.

The chronology of artefacts in this investigation adopted a “cross-dating method ([1]) ” i.e. the finds are dated by referencing datable findings recorded in published archaeological reports.

3 Site Background

3.1 South Soko

3.1.1 Geology, Topography and Hydrology

Tai A Chau is a small, hilly island with bedrock formed by fine to medium grain granite and feldsparphyric rhyolite . The hills on the island such as Fei Kei Teng, Nam Shan and Tai Chau Mei Teng range in height from 85 m to 140 m. Between the hills are valleys and small streams that run into the sea. One of the valleys was previously inhabited at an established village known as Sheung Tsuen. Valley areas have been modified by the development of agricultural fields and a fish pond, which was dug a few decades ago. In total, there are six bays around the island. These are located on the North, West and East facing shores and offer places for sheltered anchorage.

A conspicuous feature of the island is a narrow isthmus (width of approximately 40 m) which links the north and south hills of the island. This area is Site C of the archaeological survey. Prior to construction of the Detention Centre, the original ground level of the isthmus was 2.1 mPD to 6.3 mPD. The isthmus was formed of weathered soil, which overlies bedrock (rhyolite). To the east and west side of the isthmus are Tai A Chau’s two largest bays, Tung Wan and Sai Wan. The west side of the isthmus receives the most shelter during northeast, east and southeast winds due to the presence of the isthmus and surrounding hills and is therefore the most sheltered anchorage area along the island’s coast. On land, the west side of the isthmus also provided protection from winds, which may explain the formation of a second settlement, Ha Tsuen, at this location. Elsewhere on the island at the foot of the valleys and near the shore are raised beaches. (i.e. beach deposits formed when sea levels were higher than present). Sites E and G of the archaeological survey were located at these raised beach areas.

3.1.2 Archaeological Background

A group of Late Neolithic Age to Bronze Age pottery shards including an almost complete cup, a complete globular pot with impressed pattern, a complete jar and some soft net pattern pottery shards were discovered by Mr W. Schofield during a fieldwalking survey conducted between the 9^th and 12^th December 1937 at the Ha Tsuen isthmus on Tai A Chau, where the Detention Centre was later situated ([2]). Schofield, an archaeologist, compared these findings with his findings in Shek Pik Tung Wan on Lantau ([3]) and suggested that the identified artefacts on the western cliff of Ha Tsuen isthmus are proto historic period ”funeral furniture”.

Twenty years later, the Hong Kong University Archaeological Team led by S.M. Bard and J. Whitely surveyed the island on 24^th November 1957. Prehistoric chalky wares were found on the cliff face of the hillside to the north end of the isthmus (facing the beach) between Tung Wan and Sai Wan. These finds were recovered from approximately four feet (approximately 1.2m) below the ground surface. In addition, three stone adzes were discovered on the ground surface of the isthmus. It was presumed that the stone tools had been washed down from the hills ([4]).

In 1977, AMO conducted a number of visits to the island and identified three areas of archaeological interest, Spot A (the hill slope south of Pak Tso Wan), Spot B (the hill slope south of Sai Wan), and Spot C (the isthmus between Tung Wan and Sai Wan) (see Figure 3.1). Stone artefacts, pottery, many chips and polishing stones were found in Spot A and it was interpreted by AMO that the site may be a “working site”. Due to its significance, Spot A was declared as “a proposed archaeological site for the purposes of the Antiquities and Monuments Ordinance” in Hong Kong Government Gazette, No.22 for June 1983 (see Figure 3.2, upper) ([5]). A stone adze and several other stone implements were found in Spot B; and a stone adze, several prehistoric pottery shards were identified in Spot C ([6]) (see Figure 3.1).

Mr B.A.V. Peacock and Ms T.J.P. Nixon undertook further detailed archaeological survey in December 1983 by excavating seven test pits on the slope at Spot B. Substantial numbers of polishing stone tool fragments, pottery shards and archaeological features such as charcoals for heating were unearthed. In addition, a set of three complete quartz rings laying horizontally at the same level were discovered in a pit (an archaeological feature) inside a test pit within Spot B. This pit was interpreted as being for human burial from the Late Neolithic to Bronze Age. The survey recorded that the hill slopes of Spots A and B had suffered high erosion rates. Substantial cutting into the hill slope by the island’s inhabitants had revealed quantities of artefacts such as Late Neolithic to Bronze Age pottery of coarse, chalky and hard geometric types and polished quartz discs, quartz rings, stone adze fragment and whetstones ([7]). Peacock and Nixon concluded “an activity area or occupation site must have continued from the lower slope to the higher level “ ([8]) (see Figure 3.1).

In 1997-98, archaeologists from mainland China undertook another survey on the island. They conducted hand augering and surface collection. Apart from a Neolithic Age impressed net pattern pottery shard collected from the ground surface, no archaeological deposits were found ([9]). The Tai A Chau archaeological site was revised to the current boundary as shown in Figure 3.2, lower on 1^st September 1999.

During the initial stages of this EIA, in August 2004, a preliminary archaeological investigation was conducted at Spot C (known as Site C for this report) which involved excavation of four test pits on the Detention Centre platform and 20 auger holes were bored on adjacent hill slopes (see Figure 3.3). The findings are detailed in Section 4.1.3 below.

3.1.3 Historical and Ethnological Background

A review of the Xin’an County Gazetteer ([10]), the Report on Extension of The colony of Hong Kong ([11]), marine charts and old maps such as The Macau Roads of 1810, Die Chinesische Küste of 1834, the Chart of the Canton River with the Entrances & Islands of 1841 and the Map of Sun-On District of 1868 ([12]) found no record of any settlement on Tai A Chau. The first record of settlements on the island comes from the Chart of Macau to Pedro Blanco covering Hong Kong Area of 1899.

According to a stone tablet identified at a graveyard on the Island, Site E area is known as Ditangju (地塘咀). On the stone tablet, it was recorded that there were three clan groups living on the Island namely the Ng (吳) clan, Chau (周) clan and Yang (楊) clan. These clans migrated to the island from Yantian (鹽田) of Weiyang(惠陽) District in Guangdong Province in the late 18^th or early 19^th centuries during Qing dynasty. The stone tablet also indicates that the clan groups cultivated crops and went fishing for living and settled on the Island. The stone tablet also records that, in 1944 during the Japanese Occupation, the Japanese army burned down all the village houses on Tai A Chau and the whole Chau clan starved to death. The other two clans living on the Island survived and remained on the island.

An aerial photo taken in 1963 (see Figure 3.4) indicated the existence of settlements and cultivation activities on terraces. Interviews and discussions undertaken by AMO in 1977 indicated that the Sheung Tsuen and Ha Tsuen villages in Tai A Chau were established during the early 20^th century. The villagers were Hakka people who had migrated from Yantian (鹽田), north of Sha Tau Kok ([13]) in Shenzhen of Guangdong Province. At the time when this information was obtained (1977), the villages were of the 4^th generation. There were three clans in the villages, namely the Yeung clan, Ng clan and Fung clan with a total of 15 families and a population of 150. In 1981, all the private land on Tai A Chau was bought by the Hong Kong Land Co. Limited with plans for a resort development. By this time the islanders had permanently left the island ([14]). Nevertheless, buildings and structures can still be identified from aerial photos taken at the time. The population appears to have increased from the 1960s to 1981 based on aerial photo comparison. By 1981, the cultivation activities seem to have been abandoned as shown in a 1981 aerial photograph in Figure 3.5.

Peacock and Nixon recorded in their report ([15]) that there had once been a school run by a teacher from Cheung Chau Island at Ha Tsuen (1985). A reservoir (or fish pond) was found near Sheung Tsuen during their visit but it had stopped functioning as a freshwater supply by the time they visited the island (1985).

In the 1980s, a Detention Centre was constructed at Ha Tsuen. The aerial photos kept by the Lands Department show the progress of its construction over time between 1989 and 1991 and its eventual demolition in 1996. The former Ha Tsuen village had been cleared by 1989 (see Figure 3.6). This area became a construction site in 1990 ([16]), until the Detention Centre was finally completed in 1991 ([17]). In 1996 ([18]), prior to the hand over of Hong Kong back to China, the Detention Centre was cleared and all the building structures were demolished. The island is presently abandoned with no inhabitants. A Tin Hau Temple, which is still in use by local seafarers is located south west of Ha Tsuen and is the only built heritage on the Island.

3.2 Shek Pik

3.2.1 Geology, Topography and Hydrology

The bedrock of Shek Pik valley and Shek Pik Tung Wan is dominated by Mesozoic Republic Bay Formation volcanic rock and Mesozoic sedimentary rocks. These volcanic rocks mainly consist of banded rhyolitic lava, rhyolitic tuff and tuffite. These volcanic rocks have intrusions of granites, grandiorities and quartz syenites ([19]).

The geological structure of Shek Pik is dominated by northeast-southwest aligned faults, which have influenced the area’s natural drainage pattern. Run-off in stream courses generally flows in a northeast or southwest direction. The superficial deposits at Shek Pik valley are formed of Pleistocene and Holocene alluvium, which consists of pebbles or cobbles in a coarse gravelly sand matrix. These deposits have formed alluvial terraces and fans.

Shek Pik Tung Wan is a narrow, low relief, raised beach area located between the sandy shore and the base of an adjacent foothill slope. It has a ground level ranging between 4 mPD and 8 mPD and is composed of gravelly coarse reddish brown sand. The raised beach and adjacent sandy beaches both overlay a cobble substratum. To the southeast of Shek Pik Tung Wan is a small valley which has a stream.

Previous archaeological excavations carried out in the raised beach and next to the slope at Shek Pik Tung Wan revealed a layered sequence of coarse to medium sand, interspersed with remains dated to the Tang Dynasty, Bronze Age and Late Neolithic Age.

3.2.2 Archaeological Background

The Shek Pik Tung Wan Archaeological Site was first excavated by an amateur archaeologist Mr Walter Schofield and professional archaeologist Prof. J.G. Anderson from north China in March 1937 at a raised sand dune next to a stream running towards Shek Pik Tung Wan. According to the findings, they discovered prehistoric burials and two cultural layers in the sand dune dated back to early to late Bronze Age (2,800-4,000 years ago) ([20]).

In 1938, Mr. Chan Kung-chieh, a retired civil servant from Northern China excavated another site in Shek Pik. The site was Sha Gan Buey and was located adjacent to the present day Declared Monument of Shek Pik rock carving. Mr. Chan found bronze mirrors, bronze weapons, bronze knives, stone ritual objects, stone adzes, quartz rings and pottery jar fragments, which dated back to the early to late Bronze Age ([21]).

In the 1950s, an Archaeological Team of the University of Hong Kong led by Ms. M. Tregear excavated a test pit in Shek Pik. Geometric pattern shards, dating back to the Bronze Age were found.

The Hong Kong Archaeological Society and the Chinese University of Hong Kong dug 34 test pits and excavation grids at Shek Pik Tung Wan between 1979 and 1989. The Society found coarse ware shards, chalky geometric pattern ware shards, incised pattern ware shards, pebble tools and stone flakes. These dated back to the late Neolithic Age to early Bronze Ages ([22]).

The Chinese University of Hong Kong conducted an excavation at a basket ball court (a raised sand dune area) by the foothill at Shek Pik Tung Wan. According to the artefacts and soil stratigraphy, four cultural layers were identified. These dated back to the Tang Dynasty, Late Bronze Age, Late Neolithic Age. In addition a few matching rock flakes derived from chipping rocks, were found in a cobble layer, which dated back between. 3,785 B.C. to 3,200 B.C. and 4,130 B.C. to 3,640 B.C ([23]).

Shek Pik Tung Wan was also investigated as part of the first territorial wide archaeological survey conducted between 1983 and 1985. It was found the upper layer of the stratigraphy comprised a layer of rubbish of at least 1 m thick, which had been dumped at the site during the previous decade.

In the late 1990s, a second territory wide archaeological survey was conducted in Hong Kong, however, no finds were identified at Shek Pik Tung Wan ([24]) .

3.2.3 Historical and Ethnological Background

Shek Pik is one of six places on Lantau Island that made their first appearence on a published Chinese map in the late 16^th century ([25]). According to the oral history and stone tablets found at Shek Pik, there were four villages established in Shek Pik valley during the Ming Dynasty (A.D. 1368-1644). However there is no early information available on these villages except for one; by the late Ming Dynasty (the first half of the 17^th century), a walled village, Shek Pik Lo Wai, had been established ([26]). There are also records of “Shek Pik Tsuen” (石壁村) in the 1688 and 1819 editions of the Xin’an County Gazetteer (新安縣志) ([27]) and this village is also recorded in the Report on Extension of The Colony of Hong Kong dated 1898 ([28]). It is unclear whether the village name of “Shek Pik Tsuen” and “Shek Pik Lo Wai” referred to the same village or whether they are different neighbouring villages. Nevertheless, based on the review of the historic documents, it appears that village settlement existed at Shek Pik since the Ming Further support for the presence of inhabitants at Shek Pik as early as the Ming Dynasty comes from the abandoned temple at Shek Pik, the Hung Shing Temple, which was built during the Ming Dynasty (A.D. 1368-1644) ([29]).

In 1957, villages at Shek Pik were demolished and cleared to allow construction of the Shek Pik reservoir. Villagers were resettled to other places on Lantau Island and the Tsuen Wan area.

4 Archaeological Survey Findings

4.1 South Soko

According to desktop review and field investigations undertaken in 2004, six areas within the South Soko study area are considered to have archaeological potential. Fieldwork on South Soko was therefore sub-divided into six sub sites, namely Sites A to E and Site G to cover each of these areas. The following presents the fieldwork findings.

4.1.1 Site A

Site A is situated on the western side of Tai A Chau. It is a north facing hillslope area to the south of Pak Tso Wan (see Figure 4.1). The area is densely covered with trees and shrubs. The area was previously surveyed by a number of archaeologists in the late 1970s, 1982 and 1983, who found considerable quantities of artefacts such as chips and polishing stones exposed on the ground due to soil erosion or through previous human disturbance and dated to the Neolithic Age ([30]). This area was proposed to become a declared monument under the Antiquities and Monuments Ordinance in 1983 (see Figure 3.2). However, approval for the site was not given for unknown reasons. This area currently lies within the Tai A Chau Archaeological Site.

Field Walking

Since the ground surface is covered with leaves and grass, no artefacts were identified during the field walking in most of the area. However, a layer containing polished stone tools, corded pattern and geometric pattern pottery shards and pebbles was identified in a cut slope that was approximately 20m long parallel to a narrow terrace at an elevation of some 20mPD along an existing foot path (see Figure 4.2). According to the decoration pattern and stone tool manufacturing technique, the artefacts dated to Late Neolithic to Bronze Ages (some 4,000 to 2,800 years ago).

Augering

A total of 5 auger holes were bored in this area (see Figure 4.2). Two augers, A/AH4 and A/AH5, were bored near TP A1, although no artefacts were found in these two holes. The soil profile of these two holes correlates with those of TP A1. This assisted in defining the cultural layer horizontal boundary identified from TP A1 where a Middle Neolithic Age cultural layer was identified (see below for details). No cultural layers were identified for the rest of the auger holes. Details on the results of the auger holes are presented in Annex 12-C-A.

Test Pit Excavation

A total of 12 test pits were excavated (see Figure 4.2). Prehistoric artefacts were identified in three of the pits: Test Pits (TPs) A1, A2 and A3. Detailed test pit findings are presented in Annex 12-C-B. The following summarises the findings from TPs A1, A2 and A3.

The stratigraphy of the three test pits was similar, consisting of a top soil layer, subsoil layer and the weathered bed rock layer. Artefacts were found to be present in the subsoil layer (a thin colluvial layer).

Six strata were identified in these three layers at TP A1. The topsoil layer comprises L1 (greyish) and L2 (light yellow) filled soil; the subsoil/colluvial layer comprises L3 to L5 mainly in a dark brown colour; and the weathered bed rock layer comprises L6, which is a light yellow colour to reddish bed rock texture.

Three strata (L3 to L5) of TP A1 are the cultural layers. These contained 149 whitish chalky pottery shards and 205 coarse pottery shards (see Figure 4.3). Some of these coarse pottery shards had an incised wave pattern. Other pottery artefacts were an incised pattern basin and 6 reddish chalky pottery shards. Stone artefacts were also found and these included several whetstones, pebble grinders, pebble hammers and a flakes. Although a chalky ware shards were identified in L2 of TP A1, it was also found to contain modern rubbish such as plastic objects. The L2 layer is therefore regarded as filled soil layer.


Figure 4.3 Whitish Chalky Wares (Left) Coarse Pottery Shards with Incised Pattern (Right) Identified in TP A1 at Site A

By comparing the artefacts with other findings in Hong Kong and Zhuhai, it was determined that the chalky ware with stamped pattern and coarse pottery shards with incised pattern identified from L2 to L5 of TP A1, were from the Late Phase of the Middle Neolithic Age (some 4,900-5,600 years ago) ([31]).

The same kind of pottery shards and a broken polishing stone adze were identified in TPs A2 and A3.

No archaeological features were identified inside the other test pits.

Although other test pits contained no prehistoric pottery shards, one of them, TP A5, did contain a few hand-held sized pebbles. This finding indicated that the pebbles were collected by ancient people and taken to this hill slope area for some working purpose.

Summary

Based on the findings above and evaluation of natural topography, character of superficial deposit, micro-landscape, natural erosion, past land use and disturbance of the area, it is believed that an archaeological deposit exists between 0.4 m to 1.2 m below ground level between 16 mPD and 28 mPD . The horizontal extent of this deposit covers approximately 900 m² as shown in Figure 4.4. A summary of test pits findings are shown in Table 4.1.

Table 4.1 Summary of Test Pits Findings of Site A

Test Pit	Findings Result
	Prehistoric Period (Late Phase of Middle Neolithic Age, 4900-5,600 years ago)	Historic Period	No Finds
TP A1	ü
TP A2	ü
TP A3	ü
TP A4			ü
TP A5			ü
TP A6			ü
TP A7			ü
TP A8			ü
TP A9			ü
TP A10			ü
TP A11			ü
TP A12			ü

4.1.2 Site B

Site B is situated on the hill slope located south of the Detention Centre. The site is densely covered with trees and shrubs. According to the review of old aerial photos, most of the area was previously used as terraced cultivated fields in the 1950s to 1970s.

An archaeological survey of the site was previously conducted in 1983 when 12 trenches along a baseline on a gentle slope some 60 m from the Tin Hau Temple were excavated. A large quantity of artefacts including quartz ring discs, quartz cores, quartz rings, stone adzes, whetstones and corded ware shards were identified ([32]). The finding was interpreted as an “activity area” or “occupation site”. Thus, archaeological potential was considered high ([33]). ”

The area has been cut and filled for the facilities of the Detention Centre. Thus, any surviving archaeological remains at the area have been heavily disturbed or partially destroyed due to previous earth works at the area.

Field Walking

The field walking identified two artefacts, a large stone saddle-quern and a quartz disc, at a northwest facing slope at the elevation between 21mPD and 24 mPD.

The large stone saddle-quern measured 58 cm in width, 90 cm in length and 8 cm in height was made of sand stone. Both surfaces of the stone saddle-quern are smooth and flat. Five circular and smoothed areas with diameter range from 10 cm to 14 cm are identified on the upper surface (see Figure 4.5). These smoothed areas were formed due to grinding with pebble grinders. The stone saddle-quern was used for food preparation or polishing stone tools. The same kind of stone saddle-quern had been discovered in an excavation pit at Shek Pik Tung Wan in 1989 ([34]). The quartz disc measured 4.5 cm in diameter. It is white in colour and discovered in the cut slope for a terraced field (see Figure 4.5). This artefact is the waste/semi-finished material for quartz ring production. No prehistoric pottery shards were found from the field walking. The result indicated high archaeological potential at the northwest slope of Site B and thus, subsequent test pits were conducted.


Figure 4.5 Stone Saddle-quern (Left) Quartz Disc (Right) Identified in Field walking at Site B

Augering

Five auger holes were bored at Site B (see Figure 4.6). No cultural remains were found from the soil profiles of these auger holes. The details of auger hole results are presented in Annex 12-C-A.

Test Pit Excavation

A total of 16 test pits were excavated. The location of these pits focused on the northwest slope of Area B, where previous archaeological investigation previously took place. Prehistoric and historic artefacts were identified in five of the pits: TPs B1, B2, B3, B5 and B8 (see Figure 4.6). The detailed findings for the test pits are presented in Annex 12-C-B. A summary of the findings from TPs B1, B2, B3, B5 and B8, where artefacts were found, is as follows.

The stratigraphy of the five test pits are similar, comprising a top soil layer, subsoil layer and a weathered bed rock layer. The subsoil layer (colluvial layer) is the layer where cultural layers occurred.

Based on the soil texture and soil colour of TP B5, five strata were identified. The stratigraphic sequence was top soil layer (L1) → subsoil layer (L2 to L4) → the regolith layer (L5). L4 is a coarse sand layer with some gravel. In this layer were found some quartz flakes, 2 quartz cores, 3 broken quartz rings, 11 quartz discs, 2 refitting pebble cores, 1 whetstone and some corded pot shards. Furthermore, 2 archaeological features were identified (see Section 5 for details). Thus, this layer is defined as the cultural layer.

The finds from TPs B1, B2, B3, B5 and B8 are classified into two categories, artefacts and archaeological features. The artefacts are prehistoric pottery shards, pebble grinder, whetstone, pebble hammer, quartz flakes, quartz core, quartz discs, broken quartz rings. According to their typology and manufacturing techniques and a comparison with same kind of finds in Hong Kong and the Pearl River Delta, they date back to the early phase of the Late Neolithic Age some 3,500 to 4,400 years ago ([35]).

Three archaeological features were discovered in two test pits, B1 and B5. One of the features recorded as Feature 1 (F1) was identified in L4/F1 of TP B5. It consists of a cluster of pebbles and cobbles, a few coarse ware shards, 1 broken quartz ring, whetstone, 2 quartz cores and 1 quartz ring disc. At Feature F1 there were two pieces of a broken pebble. These pieces could be fitted together to form one large pebble though a small part was missing (see Figure 4.7). The position of this feature was close to another feature F2 of B5 (see Figure 4.7).

The other feature (F2) comprised a cluster of 10 quartz ring discs behind a large cobble in L4 of TP B5 (see Figure 4.7). The discs were arranged in an orderly manner indicating that they are an undisturbed in situ deposit.

The third feature (F3) is a burial pit found in L2 of B1. It is a circular pit with 0.96 m diameter and a depth of 0.8 m. A brown cracked glazed burial pot with two layers of lids was identified in the pit (see Figure 4.8). A human skeleton was found in the pot. Accordingly to the glaze and shape of the pot, it dates from the late Ming to Middle Qing Dynasties (the 17^th to 18^th centuries). This kind of pot is seldom reported in published archaeological papers in Hong Kong and Guangdong region.

Summary

Based on the findings as mentioned above and evaluation of natural topography, character of superficial deposit, micro-landscape, natural erosion, past land use and disturbance of the area, it is believed that an archaeological deposit occurs between 0.3 m and 0.8 m below ground level at between 17 mPD to 30 mPD at Site B. The horizontal extent covers approximately 800 m² as shown in Figure 4.4. The summary of test pits findings are shown in Table 4.2.

Table 4.2 Summary of Test Pits Findings

Test Pit		Findings Result
	Prehistoric Age (Late Phase of Late Neolithic Age, 3,500 -4,400 years ago)	Historic Age (Late Ming to Middle Qing Dynasties, or 17^th to 18^th centuries)	No Finds
TP B1	ü
TP B2	ü	ü
TP B3	ü
TP B4			ü
TP B5	ü
TP B6			ü
TP B7			ü
TP B8	ü
TP B9			ü
TP B10			ü
TP B11			ü
TP B12			ü
TP B13			ü
TP B14			ü
TP B15			ü
TP B16			ü

4.1.3 Site C

Site C is situated at the Detention Centre site. Prior to the construction of the Detention Centre, the western portion of this site was the location of a village known as Ha Tsuen (Lower Village), which faces Sai Wan.

Late Neolithic Age to Bronze Age pottery shards had been identified in this area in 2004. However, the area has been heavily modified for the construction and demolition of the Detention Centre during the late 1980s to the early 1990s. Thus, it is considered Site C has high archaeological potential but has been severely disturbed by the development of the Detention Centre.

Survey Findings for 2004 Archaeological Investigation

During the initial stages of this EIA, in August 2004, a preliminary archaeological investigation was conducted at the preliminary proposed Project Area boundary (PPA) as shown in Figure 2.1.

Field walking was undertaken on the accessible hill areas and coastal area within the PPA. Only three historic period pottery shards were found on the surface of hill slopes to the south and north of the Detention Centre. These finds were two Qing Dynasty (A.D.1644-1911) pot and bowl shards collected on the south hill slope of the Detention Centre between TP3 and TP4, and a Song to Yuan Dynasty (A.D.960-1368) cracked celadon (a kind of light green glaze) bowl rim found besides a footpath on the hill slope to the north of the Detention Centre (see Figure 4.9).

Figure 4.9 Song to Qing Dynasties Pot and Bowl Shards Discovered at Both Hillslopes next to the Detention Centre Site

Many shells were exposed on both hill slopes to the north and south of the Detention Centre. It is likely that the users of the Detention Centre left them behind.

A total of 20 auger holes (AH) were bored (see Figure 3.3 for locations and detailed recording in Annex 12-C-C). The Detention Centre site is presently covered with a concrete layer and is not suitable for hand augering. Consequently, augering was undertaken on the two hill slopes to the north and south of the Detention Centre. The findings indicated that the topsoil layer consisted of brown sandy soil commonly found within the first 20cm from the ground, this soil profile regarded as O and A horizons. Beneath the topsoil layer is a yellowish sandy soil layer on top of the bedrocks or its debris, this soil profile is a subsoil layer regarded as B to C horizons. Each auger hole had a diameter of approximately 10 cm and reached a maximum depth of 0.2 to 1.3 m depth below the existing ground surface. The stratum of the soil and deposits were recorded and are detailed in Annex 12-C-C.

No distinctive compact dark coloured stratum (if this contains artefacts it is regarded as a cultural layer) was identified from augering. This suggests that no cultural layer exists at the hill slope areas. All hand auger holes reached the C horizon. No cultural layer or ancient cultural remains were found in the auger holes.

Four test pits (TP) were excavated within the PPA (see Figure 3.3). Detailed strata descriptions on the TPs are presented in Annex 12-C-C.

Two test pits, TP1 and TP2, measuring 3 m x 3 m and 2 m x 3 m, both at an elevation of 6.71 mPD, were dug at the Detention Centre (see Figure 3.3). When correlated the stratigraphy between C/TP1 and C/TP2 at the Detention Centre site, they are of almost the same stratum sequence as follows (see details in Annex 12-C-C):

Concrete Floor® Filled soil ® Original topsoil layer ® Late Bronze Age Layer ® Late Neolithic Age Layer®Sterile Layer ® Bed Rock Textural Debris Layer.

In Layer 6 of C/TP2, regarded as Late Bronze Age layer, artefacts such as a few pottery pot shards, hand-held size pebbles and a pumice were revealed (see Figures 4.10, 4.11 and 4.12).

One of the shards was a typical datable decoration pattern, an impressed “double f (kui 夔) ” pattern (Figure 4.11). This pattern of pottery is commonly found in Hong Kong and both Guangdong and eastern Guangxi Provinces. In addition, two hand-held size stone tools and two coarse ware pot rims were identified. Those stone tools are like-cleaver tools for either chopping or cutting purposes, and a pebble with artificially flattened surface for grinding.

Figure 4.10 Hand-held Size Pebble and Working Stone from C/TP2 L6

Figure 4.11 Double-f Pattern Pot Shard from C/TP2 L6

Figure 4.12 Coarse shards from C/TP6 L6

A piece of chalky pot shard without decoration was unearthed in Layer 7 of C/TP2. According to its texture, it is suggested that the shard is dated to Late Neolithic Age ([36]).

According to the desktop review, it is understood that C/TP1 and C/TP2 were excavated at the isthmus where the former Ha Tsuen village was situated. The finds indicated that the isthmus area was suitable for human habitation. A pumice (see Figure 4.13) is ancient environmental evidence that was transported to the shore by sea waves, the discovery of pumice and pebbles in Layer 6 in TP2 indicated that the layer was the original ground level of some 3,000 years ago. The recovery of the pottery shard with a decorated pattern double f commonly found in Late Bronze Age pottery wares conforms with the “hard geometric” pattern pottery identified from previous archaeological investigations since 1937. This result implies that there was ancient human inhabitation on Tai A Chau since 3,000 years ago. The result also suggests that at least one stable ancient cultural layer from the Bronze Age still survives beneath the Detention Centre site at the isthmus at an elevation of 4.7 mPD to 5.8 mPD (the existing ground level at the range of 6.7 mPD to 6.8 mPD).

Figure 4.13 Pumice from C/TP6 L6

Two test pits, C/TP3 and C/TP4, measuring 1 m x 1.5 m at elevation from 24.55 mPD to 25.39 mPD at the hill slope to the south of the Detention Centre (see Figure 3.3). The strata of C/TP3 and C/TP4 are similar. The ground surface is covered with shrubs and overgrown grass. Yellow red to reddish brown coarse sandy soil were regarded as subsoil layers beneath the topsoil layer. Boulders measured as large as 70 cm in length appeared throughout all layers, especially in subsoil layers both at C/TP3 and Layer 4 in C/TP4. This suggested that the strata are natural deposit without human disturbance. No artefacts of archaeological significance were discovered in C/TP3 and C/TP4, which indicates no cultural layer (primary deposit) on the southern slope of Detention Centre.

In summary, two cultural layers dated to Late Neolithic and Bronze Ages were identified as surviving beneath the concrete layer of the Detention Centre at the original isthmus between Sai Wan and Tung Wan on Tai A Chau. It is believed that the original isthmus area is at an elevation between 4.67 mPD and 5.8 mPD covered by the concrete platform (see Figure 4.14)

Survey Findings for 2005 Archaeological Investigation

Based on the 2004 findings result, further archaeological investigation was undertaken in 2005 focus on test pitting to define the horizontal extent of the deposits.

As Site C had been heavily modified by previous site formation works and covered with concrete, no fieldwalking and augering was undertaken. Test pit allocation was based on the analysis of 1960s to the 1990s aerial photos and old maps to establish the original landform to demarcate possible extent of the original isthmus.

A total of 6 additional test pits were excavated (see Figure 4.15) adjacent to the possible original boundary of the isthmus to confirm the deposit boundary (see Figure 4.14). The deepest pit was excavated down to 3.5m below ground level. Details of the test pit results are presented in Annex 12-C-B. Summary of test pit findings are described below.

The stratigraphic sequence of TP C6 consists of a surface concrete floor (L1)→ filled soil layer (L2 to L7)→ first phase Detention Centre concrete floor (L8, approximately dated to 1988)→filled soil layer (L9)→original village concrete floor (L10, approximately 1970s) →isthmus sandy soil (L11-14).

The test pit was dug to the L11 layer and the sequence below this level was determined by boring an auger hole down to a depth of 1.5 m that reached +1 mPD. No artefacts were found in any of the layers of the test pit.

The first phase Detention Centre concrete floor was found in the stratigraphic profile of all other pits except TP C3. In TP C3, granite bedrock was found beneath first Detention Centre concrete floor layer.

Although TPs C1 and C4 were not excavated as deep as C6, the stratigraphy of these test pits was similar. This indicated that TPs C1, C4 and C6 have the same geological layers. The findings supported the desktop review and confirmed that the area at TPs C1, C4 and C6 is outside the western boundary of the archaeological deposit area.

No cultural deposits were yielded from TP C2. This result indicates that the archaeological deposits did not extend as far as this test pit’s location. TP C2 is outside the northern boundary of the archaeological deposit area. Similarly, no cultural layers were identified in TPs C3 and C5. This indicates these test pits are located to the east of the archaeological deposit boundary.

Summary

Based on the findings of the surveys undertaken in both 2004 and 2005 and evaluation of natural topography, character of superficial deposit, micro-landscape, natural erosion, past land use and disturbance of the area, the extent of the Site C archaeological deposit has been demarcated. The findings from this excavation provide further support and confirmation of the extent of the archaeological deposits, which is shown in Figure 4.4. The archaeological deposits cover a total area of approximately 1,600 m²(40 m x 40 m). Within this area, there is a core of archaeological deposits, which is correlated to the original isthmus. With reference to the original isthmus and elevation, the core of the archaeological deposits is estimated to be 800 m²(20 m x 40 m). The summary of test pits findings are presented in Table 4.3.

Table 4.3 Summary of Test Pits Findings

Test Pit		Findings Result
	Prehistoric Age	Historic Age	No Finds
TP C1			ü
TP C2			ü
TP C3			ü
TP C4			ü
TP C5			ü
TP C6			ü

4.1.4 Site D

Site D is located on a natural coastal terrace on the northern side of the east bay, Tung Wan. Prior to development of the Detention Centre, the adjacent shore was a sandy beach. Currently, the shore consists of a granite block artificial seawall. No archaeological survey has previously been undertaken in this area. This site was investigated because it was considered that the original topography of this coast may have been a suitable location for human activities. Thus, it is considered to have archaeological potential.

Field walking

Field walking was conducted at the terrace and along the existing footpath onto the steep hill slope area at the back of the terrace. No finds were identified.

Augering

Six auger holes were bored at the terrace area (see Figure 4.16). Auger holes D/AH 5 and D/AH6 which were bored on the terrace revealed a dark brown stratum composed of sandy soil. No artefacts were identified from the two auger holes on the terrace or from auger holes made on the adjoining slope. Nevertheless, it was noted that the dark brown layer found on the terrace is identical to the dark brown layer of TP D1 where cultural remains were identified. This assisted in defining the horizontal extent of the archaeological deposit. Details of the auger hole results are presented in Annex 12-C-A.

Test Pit Excavation

Three TPs, test pits D1, D2 and D3, were excavated at Site D (see Figure 4.16). Details of the test pits results are presented in Annex 12-C-B. A summary of test pit findings are described below.

Of the three test pits, artefacts were found only in one, TP D1. The surface layers of TP D1 (L1 and L2) consisted of filled soil, mixed with modern construction waste and rubbish. Below these, layers L3 and L4 are regarding as a natural stratum. Artefacts were found in the deeper L5 layer, which consisted of dark brown sandy soil. In this layer were found historic and prehistoric artefacts. These included remains from lime kiln structures artefacts dated to Tang Dynasty (A.D. 618-907), celadon bowl shards dated to Song Dynasty (A.D. 960-1279) and a lot of coarse corded pot shards and rims, chalky geometric pattern pot shards, and stone tools (scraper, flake-knife) and a lot of flakes (see Figures 4.17 and 4.18). According to the pottery fabric and its decoration patterns, they date back to the Late Neolithic (3,500 to 4,400 years ago) and Bronze Ages (2,800 to 3,500 years ago) ([37]).


*Figure 4.17 Corded Pottery Rim (Left) Stone flakes (Right) Identified in L5 of TP D1 at Site D*


*Figure 4.18 Flake-Knife (Left), Geometric Pattern Pottery Shards (Right) Identified in L5 of TP*D1

Summary

The findings indicated that archaeological deposits at Site D occur some 0.8 m to 1.3 m below ground level at approximately 8 mPD. In addition to the evaluation of natural topography, character of superficial deposit, micro-landscape, natural erosion, past land use and disturbance from the construction and decommissioning of the Detention Centre at Site D, it is believed that the horizontal extent of the archaeological deposits cover an area of approximately 100 m²(20 m x 5 m) as shown in Figure 4.4. A summary of test pit findings is presented in Table 4.4.

Table 4.4 Summary of Test Pits Findings

Test Pit		Findings Result
	Prehistoric Age (Late Neolithic Age to Early Bronze Ages, 2,800 to 4,400years ago)	Historic Age (Tang and Song Dynasties, A.D. 608-1279)	No Finds
TP D1	ü	ü
TP D2			ü
TP D3			ü

4.1.5 Site E

Area E is situated at the eastern part of Tai A Chau. It comprises a slope area, a small valley with a north running stream, terraced fields and a sandy raised beach facing northeast. According to local people, the bay is known as San Shek Wan (散石灣). A graveyard is located on the hill slope area. According to the inscription texts of grave headstones, the place is known as Ditangju (地塘咀). The area was previously surveyed by an archaeological team from Guangdong Province in 1997. Only a few 20^th century blue-and-white shards were identified from a test pit excavated at the beach area ([38]).

Desktop review indicated that the area presently occupied by abandoned terrace fields and the gentle slope area of the valley were suitable locations for ancient activities. Thus, it is considered to have archaeological potential.

Field Walking

Field walking was conducted across the terraced fields and gentle slope areas. Due to dense vegetation cover, no artefacts were identified.

Augering

A total of 15 auger holes were made in the valley and slope area (see Figure 4.19). The soil profile at the valley bottom comprised dark grey mud indicated that this area is a wetland. The soil profile from the slope area consisted of dark brown to yellow sandy soil and loamy soil which is indicative of colluvial soil. In addition, a layer of modern construction waste was identified at the terrace area at the southern part of Site E. No artefacts were identified from the augering. Details of the augering results are presented in Annex 12-C-A.

Test Pit Excavation

A total of 17 test pits were excavated in various landform areas. Prehistoric and historic artefacts and ecofacts were discovered in 3 of the 16 test pits.

There are four landforms at Area E, comprising the raised beach (sand dune), hill slope, coastal terraces and valley. In order to verify archaeological potential, test pits were allocated to different landforms. TP E1 is located at the raised beach and TPs E2, E3, E4, E16 and E17 are allocated at the coastal terraces. In the same way, TPs E5, E11 to E15 is located at the hill slope terrace, TPs E7 to E9 at the hill slope and TPs E6 and E10 at the valley. Details of the test pit results are presented in Annex 12-C-B. A summary of test pit findings are described below.

TP E1 was excavated at a raised beach with an elevation of 5.6 mPD (see Figure 4.19). Based on the soil texture and colour, the stratigraphy was divided into six layers, L1 is a top soil layer, L2 to L5 is sand layer and L6 is a cobble layer.

A large quantity of ecofacts (including shells, fish crania, fish jawbone, mammal tooth and animal bones) and artefacts (pot shards, stone flakes, flake-knife, broken blades, scraper and fired clays) were discovered in L3. This layer comprises dark yellow coarse sand. Since ecofacts associated with artefacts were found but no structural remains such as house floor or post holes were found, it is believed that TP E1 is a dumping area which should not be far away from the living area, probably on the hill slope terrace.

A total of 1,803 shells were found in L3 of E1. The shells mainly came from the marine snails Nerita albicilla, Monodonta labio, Chlorostoma argyrostoma, Thais luteostoma, and the limpet Cellana grata. In addition, there are 206 fish bones identified and one of the species identified is Pomadasys hasta (head grunt) (see Figures 4.20 and 4.21) ([39]) .

A modern construction waste layer was identified in both TPs E16 and E17 on the coastal terrace. This layer of dumped rubbish is solid evidence that the original ground surface soil or even the subsurface soil had been modified or disturbed.


*Figure 4.20 Fish Bones (Left)* Head Grunt Head Bones (Right) *Identified* *in L3 of TP E1* *at Site E*

Figure 4.21 Fish (Grunt) Identified in L3 of TP E1 at Site E

Pottery artefacts found at TP E1 included 150 pot shards, corded coarse ware pot shards and some shards with raised square and triangle patterns (one kind of geometric design) impressed on the exterior (see Figure 4.22). Based on the decoration pattern and shards fabric, L3 of TP E1 is dated to the Bronze Age (about 2,400 to 3,500 years ago), also known as the Middle to Late Western Zhou to Spring and Autumn Periods in China.

Figure 4.22 Raised Square(right) and Net Pattern(left) Pottery Shards Identified in L3 of TP E1

TP E2 is located at the coastal terrace area. The stratigraphy is that L1 is the topsoil, L2 and L3 are subsoil and L4 is weathered bed rock. A Song Dynasty celadon bowl shard and a rim were identified in L2 and L3.

TPs E5 to E9 were dug on terraces and hill slopes on the west side of the valley. The stratigraphic sequence of these test pits are almost the same, comprising top soil subsoil and then weathered bed rock layer.

No artefacts were found in TPs E3 to E7 and TPs E9 and E10. A piece of prehistoric coarse ware shard was found in L2 of TP E8 indicating that archaeological potential exists at the western gentle slope area at Site E.

TPs E6 and E10 were dug on the northern end of the valley in Site E. The E6 test pit located on the hill slope was observed to be composed of colluvial deposit. Dark grey mud was observed in TP E10.

TPs E11 to E15 were excavated at the slope area, while TPs E16 and E17 were excavated at a coastal terrace at the southern part of Site E. No artefacts or archaeological feature were discovered in these test pits. These test pits were composed of colluvial soil. In addition, a layer of modern construction waste was observed in TPs E16 and 17, which indicated that the terrace area had been used as a rubbish dump.

Summary

The findings indicated that three archaeological deposit areas were identified in Site E. One of these is located at the raised beach area, while the others are located at the coastal terrace and valley gentle slope. To demarcate the archaeological deposit area of Site E, the natural topography, character of superficial deposit, micro-landscape, natural erosion, past land use and disturbance of the area was evaluated. The evaluation concluded that the extent of the archaeological deposit is 8 m x 10 m at the raised beach area, 25 m x 10 m at the coastal terrace and 8 m x 15 m at the gentle slope.

The archaeological deposits are located 0.4 m to 0.9 m below ground level at both the raised beach and coastal terrace areas and 0.3 to 0.5 m below ground level at the gentle slope area. The archaeological deposits cover areas of approximately 80 m²at E1, 250 m² at E2 and 120 m² at E3 respectively as shown in Figure 4.4. The summary of test pits findings are presented in Table 4.5.

Table 4.5 Summary of Test Pits Findings

Test Pit	Findings Result
	Prehistoric Age Late Neolithic Age to Bronze Age (2,250 to 4,000 years ago)	Historic Age Song Dynasty (A.D. 960-1279.) and late Qing Dynasty	No Finds
TP E1	ü
TP E2		ü
TP E3			ü
TP E4			ü
TP E5			ü
TP E6			ü
TP E7			ü
TP E8	ü
TP E9			ü
TP E10			ü
TP E11			ü
TP E12			ü
TP E13			ü
TP E14			ü
TP E15			ü
TP E16			ü
TP E17			ü

4.2 Site G

Site G is located next to a bay (Pak Tso Wan) to the northwest of Nam Shan (see Figure 4.4). The site is mainly covered with trees and shrubs. Previous archaeological investigation had been conducted in the 1990s at this area but nothing was found.

There are four types of natural sediment deposits in the valley. The modern sand dunes consist of fine sand. The wet agricultural land consists of grey mud and whitish and yellowish fine sand. The lower valley is composed of dark grey mud and upper valley has deposits of brown to yellow colluvial soil.

Field Walking

Field walking was undertaken on the beach, abandoned fields and the valley. Due to dense vegetation cover, no artefacts were identified from field walking.

Augering

A total of 4 auger holes were bored within Site G (see Figure 4.23). Although no cultural remains were identified from augering (see Annex 12-C-A for details), Bronze Age shards had been revealed at the interface between Layers 3 and 4 of TP G2. As there is a correlation of stratigraphy between AH G1 and TP G2 (AH G1 is located 12 m north of TP G2), this indicated that AH G1 is in an area of archaeological potential and that the archaeological remains covers area where AH G1 located.

Test Pit Excavation

A total of 3 test pits were excavated at Site G (see Figure 4.23). Two of the test pits (TP G1 and G2) were allocated to the lower valley with alluvial sediment The other test pit (TP G3) was dug at the upper valley. This allowed differences in the stratigraphy of the test pits located in the upper and lower valley to be identified.

The stratigraphic sequence of TPs G1 and G2 in the lower valley area is top soil layer →grey sand and dark bluish gray → yellow sand→cobble layer. This stratigraphy was indicative of the existence of a past lacustrine (lagoon) environment. Some artefacts dated to Bronze Age and Song and Yuan Dynasties (A.D. 960-1368) were found in test pit TP G2. The presence of these ancient artefacts which were deposited in lagoon sediments as well as their eroded rounded edges suggests that they were washed down from the slope nearby or from the flat area of upper valley. The stratigraphy of TP G3 at the upper valley is top soil→subsoil layers →regolith layer. (see Annex 12-C-B for details)

Artefacts comprising pottery shards and a stone tool were discovered in only one of the three test pits excavated, namely TP G2, which was excavated in the lower valley at an elevation of 4.63 mPD. In this pit, pottery was found at the interface between two strata, L3 and L4. The pottery artefacts consisted of 1 piece of light green crackled glazed rim shard, 1 piece of dot with square pattern of coarse sand shard, 2 pieces of plain whitish chalky shards and 2 pieces of corded as well as 57 plain coarse ware shards (5 pieces of pot rim). Based on the glaze character of the shards, they date back to the Song to Yuan Dynasties and Bronze Age. A stone chopping tool made of coarse tuff was also discovered amongst the Bronze Age pottery shards at the interface between L3 and L4 of TP G2.

Summary

Based on the fieldwork findings, archaeological deposits occur in the lower valley area of Site G as shown in Figure 4.4. At the lower valley, archaeological deposit is present between 0.3m and 0.6m below ground level. The artefacts that were found were located in the stratum consisting of alluvial grey and yellow sand. The pottery shards were observed have rounded edges, which indicates that these artefacts are a secondary deposit. With reference to the drainage pattern of the area, the primary deposit is located further up the slope at the upper valley higher than the Site G known archaeological deposit area.

To demarcate the archaeological deposit area of Site G, the natural topography, character of superficial deposit, micro-landscape, natural erosion, past land use and disturbance of the area had been evaluated, it is believed that there are archaeological deposits that cover areas beyond where TPG2 and AH G1 are located. It is estimated that the archaeological potential area covers approximately 600 m²(see Figure 4.4).

The summary of test pits findings is presented in Table 4.6.

Table 4.6 Summary of Test Pits Findings at Site G

Test Pit		Findings Result
	Prehistoric Age (Bronze Age, 2,300 B.C. to 250B.C.)	Historic Period Song and Yuan Dynasties (A.D.960-1368)	No Finds
TP G1			ü
TP G2	ü	ü
TP G3			ü

4.3 Shek Pik

The Site (Site F) is situated on an area spanning low terrain and adjoining lower hill slopes close to the coast from the east to the north side of the bay called Tung Wan at Shek Pik on Lantau Island. The site is mainly covered with trees and shrubs. Archaeological excavations have previously been conducted in this area and it is a known archaeological site called the Shek Pik Tung Wan Archaeological Site (see Figure 4.24).

Field Walking

The ground surface within the Shek Pik Tung Wan Archaeological Site is currently filled with a layer of over 1m of 1970s rubbish. A few lime kiln structures remain, such as broken bricks and fired bars which were identified in an area next to an electric post (see Figure 4.25).

Figure 4.25 Lime Kiln Structure Remains Identified from Field Walking at Shek Pik

Augering

A total of 19 auger holes were bored within the Shek Pik Study Area (see Figure 4.26). Nine of the auger holes were located on the low terrain on the east side of Tung Wan. Four others were bored in the vicinity of the rock carving (a declared monument) and the remaining six were bored to the south of the Country Park Office close to the coast on the north side of the bay. No cultural remains were found during the augering (see Annex 12-C-A for details).

Test Pit

A total of seven test pits were excavated in the Shek Pik Study Area. Five of these were excavated on the low coastal area on the eastern side Tung Wan bay, while the other two were excavated in an area to the south of the Country Park Office on the northern edge of Tung Wan bay (see Figure 4.26). Of the five test pits dug on the east side of Tung Wan, three (SP1, SP3 and SP5) were located on the raised beach, while the other two (SP 2 and SP4) were located on the adjacent hill slope area. In this way, the test pits encompassed areas with different stratigraphies arising due to different landforms at the site.

It was found that the stratigraphic sequence in test pits located on the hillslope area goes from top soil layer to subsoil layer and subsequently down to a regolith layer (weathered bed rock layer). In contrast, the stratigraphy of test pits located on the raised beach is topsoil→sand layer →cobble layer.

In terms of archaeological finds, ancient artefacts were only found in two of the five excavated test pits. The test pits with artefacts were TP SP1 and TP SP4.

Excavations for TP SP1 were carried out on the raised beach at an elevation of 7.4 mPD. The pit was excavated to 5.3 mPD, which was close to the bottom level of the test pit excavated in 1979, 1987, 1988 and 1989 by the Hong Kong Archaeological Society and The Chinese University of Hong Kong. In this pit, three strata were found to contain artefacts . According to the typology and character of the finds, they are dated to Tang Dynasty and Late Neolithic Age.

The stratigraphy of TP SP1 was composed of 6 layers. Layers 1 to 4 of SP1 mainly comprise sandy soil and some waste material, which indicate that these layers have been disturbed by previous activities. Lime kiln structure remains such as broken bricks were found in L3 that dates back to the Tang Dynasty

Pebbles identified in L4 and a stone flake with a few striking scars were identified in L5 with compacted dark yellow coarse soil. The stratigraphic sequence of the survey’s test pits was found to correlate with previous 1987-1988 excavations([40]) ([41]) ([42]). The comparison indicated that Layer 5 of SP1 is identical to Layer 9 of the 1988 -1989 excavations. According to absolute dating results, this Layer 9 was deposited between BC3,200 to BC3,785([43]). Artefacts in L4 and L5 are therefore considered to be from the Late Neolithic Age. Below this layer, L6 comprises a cobble layer from which no artefacts were found.

The other test pit which contained artefacts was TP SP4, located at the foothill slope on the east side of Tung Wan bay. The stratigraphic sequence of SP4 was top soil (layer L1) → subsoil of yellow coarse sandy soil (layer L2) → regolith (layer L3) of whitish to yellow weathered bedrock texture soil. A small piece of coarse ware shard was found at L2. Based on its fabric, the shard was dated to the Late Neolithic Age some 4,000 years ago.

Summary

Taking into account the survey findings and a review of previous archaeological findings from this area, it is considered that archaeological deposits occur at the raised beach area and the gentle hill slope area as shown in Figure 4.27. To demarcate the archaeological deposit area of Shek Pik (Site F), natural topography, character of superficial sediment, micro-landscape, and disturbance of past land use were not only taken into account, but also correlation of test pits and augers hole soil profile and referred to pervious archaeological surveys, these archaeological deposit of Shek Pik cover an area of approximately 1,500 m². At the hill slope area, archaeological deposits exist between 0.4m to 0.7m below ground surface at an elevation of 9 mPD to 13 mPD. At the raised beach area, archaeological deposits exist between 0.5 m to 2.3 m below a ground level of 7.4 mPD. A summary of the test pits findings is presented in Table 4.6.

Table 4.6 Summary of Test Pits Findings in Shek Pik

Test Pit	Prehistoric Age (Late Neolithic Age, 4,000-3,400 years ago) ~~Prehistoric Age~~ ~~(Late Neolithic Age, 4,000-3,400 years ago)~~	Findings Result Historic Period	No Finds
TP SP1	ü
TP SP2	(		ü
TP SP3	(		ü
TP SP4	ü
TP SP5			ü
TP SP6			ü
TP SP7			ü

5 Preliminary Finds Assessment

5.1 South Soko

5.1.1 Artefacts and Ecofacts

The artefacts unearthed at South Soko comprised shards of pottery and porcelain, stone tools, stone ornaments, stone ornament manufacture wastes and flakes. Ecofacts found during the excavations were shells, fish bones, animal bones and an animal tooth. The artefacts and ecofacts were dated based on their position in the stratigraphic sequence, by using cross dating comparison techniques and by correlating with absolute dating records from the literature. The artefacts and ecofacts discovered in South Soko Island during the surveys undertaken between 2004 and 2006 are presented in Table 5.1. Special finds are further detailed below.

Table 5.1 Summary of Artefact and Ecofacts Found for the Surveys Understaken between 2004 and 2006

Test Pits	Layer	Pre-historic pottery	Stone tool	Stone flake & core	Quartz ring	Quartz disc	Song celadon	Blue and White Porcelain	Modern depsoit	Ecofact
C/ TP2	L6	1(dof)	Some pebbles
	L7	1(ch)
A1	L2	6(cor)
	L3	42(ch), 21(cor)	1(h), 3(pe)	8(f)
	L4	22(cor), 16(ch)	1(h), 2(g), 3 (pit)	1(f)
	L5	106(ch), 83(cor)	4(0), 1(u)
A2	L2	36(cor), 2(ch)	1(g)	1(f)
A3	L2	50(cor)	1(g), 1(ad), 2(u)	1 (f)
B1	L2	4(cor)		3(f)
	L3	1(cor)	1(u)	2(f), 4 (r )
B2	L3			1(f), 1(c)
B5	L3	33(cor)		7(f)
	L4	57(cor), 7(ch)	4(u), 3(g),	7(f), 6 (c )	1	10
B8	L2				1				10
B12	L2
C6	L9							7	21	4(sh)
D1	L5	84(cor), 12(geo)	1(kn)	13 (r )						4(sh)
E1	L1	2(cor),						1
	L2	1(cor)
	L3	23(cor), 117(geo)	2(u)	87(r ), 9(f)						1(te), 79(b), 1803 (sh)
E2	L2						2		1(p), 15(rf)
	L3		1(g)				1		2(rf)
E4	L1								6
E7	L1	1(cor)
E8	L1								4
G2	L2	59(cor), 2(ch), 1(geo)		1(cho)			1

Note: ad=adze, b=bone, c=core, ch=chalky shard, cho=chopping tool, cor=coarse shard, dof=double f pattern shard, f=flake, g=grinding stone and whet stone, geo=geometric shard, h=hammer , kn=knife, p=pounder, pit=pitted pebble, pe=pebble, r=stone raw material, sh=shell, te=animal tooth, u=function unclassified.

5.1.2 Pottery

A total of 783 prehistoric pottery shards and 4 Song Dynasty celadon bowl shards were found in South Soko Island for the surveys undertaken between 2004 and 2006. Based on the decoration pattern, texture and functions, pottery shard artefacts were classified into three main types: coarse ware, chalky ware and porcelain. Different pottery shards were dated as belonging to the Late Phase of Middle Neolithic Age (4,000B.C.-2,500B.C.), Late Neolithic (2,500B.C.-1,500B.C.), Bronze Ages (1,500B.C.-221B.C.), Tang, Song, Ming and Late Qing Dynasties (A.D. 608-1911). Table 5.2 summarises the pottery shards found according to their dating.

Table 5.2 Pottery Shards found on South Soko Island between 2004 and 2006

Period	Middle Neolithic Age			Late Neolithic Age	Bronze Age		Song Dynasty	Late Ming to Late Qing Dynasty
	4,000 B.C.– 2,500 B.C.			2,500 B.C.– 1,500 B.C.	1,2000 B.C.– 220 B.C.		A.D. 608 – 1911
Pottery Type	Plain Whitish Chalky Shards	Incised and plain Reddish Chalky Shards	Corded and Incised Coarse Ware Shards	Coarse Corded Ware Shards and Chalky Shards	Geometric, Plain, Double-f Pattern, Corded ,Coarse Shards and Plain Chalky shards	Hard Geometric Shards	Celadon Bowl Shards	Blue-and white Porcelain bowl shards and dark brown glazed shards
Site A	160	6	218
B				102				2
C					2			28
D				78	12	1	2	16
E				27	117			6
G					61		1

Coarse Ware Shards

The coarse ware shards were dated to the Late Phase of the Middle Neolithic, Late Neolithic and Bronze Ages. A total of 218 coarse ware shards were found in three TPs: TP A1, A2 and A3 at Site A. The majority of the shards have corded and incised wave patterns on their surface (see Figure 5.1). In addition, a few shards have incised wave designs both on the upper neck and interior surface of the rim or incising horizontal lines over the corded pattern. This is a typical design of globular pot dated to the Late Phase of Middle Neolithic Age. Elsewhere in Hong Kong, the same kind of finds have previously been unearthed at the Sham Wan Site on Lamma Island ([44]) and Kwo Lo Wan Upper Site on Chek Lap Kok Island ([45]). Coarse ware pottery shards were found at Site B, Site D and Site G.

Figure 5.1 Corded Pattern Coarse Ware Shards Identified in TPA1

In test pits B1 and B5 at Site B, a total of 95 coarse ware shards, which have an impressed corded or plain pattern were found. At Site D, 84 coarse ware shards with a corded pattern were found in test pit D1. The shards from these two sites were dated to the Late Neolithic Age (some 3,400-4,000 years ago) (see Figure 5.2).

Figure 5.2 Corded Pattern Coarse Ware Rim Shards Identified in L5 of TP D1

At Site G, a total of 59 coarse ware shards were found in TP G2. The majority of these have corded patterns (see Figure 5.3(a), lower row, right and left), while some were plain with no decoration. One of the coarse ware shards, which was found in the same layer as the others, has a dotted square pattern (see Figure 5.3(a), lower row, middle). This indicates that the layer where coarse ware shards were found dates to the Bronze Age.

Geometric Design Coarse Ware Shards

Coarse ware shards with geometric design patterns occurred at two sites, Sites E and G. In total, 118 pieces of geometric design coarse ware shards were unearthed in TP E1 of Site E (see Figure 5.4). Elsewhere in Hong Kong, similar coarse ware shards impressed with this geometric design pattern were previously found at the Tung Wan Tsai Site on Ma Wan Island. Using absolute dating techniques, the shards on Ma Wan Island date from 3,210 to 3,630 years ago ([46]).

Figure 5.4 Geometric Design Coarse Ware Shards Identified in L3 of TP E1

As mentioned previously, at Site G, a single shard piece was impressed with dots in square pattern. This is a typical design of globular pots (see Figure 5.3(a), lower row, middle) in the Bronze Age (Middle to Late Zhou Dynasty to Autumn and Spring Period) in the Hong Kong region. Therefore, this shard, dates back some 2,200 to 3,500 years. Similar type of finds have not only been found in Hong Kong but also in the northern Guangdong Province area ([47])([48])([49]).

Chalky Ware shards

Chalky ware shards were unearthed at Sites A and G at South Soko. The chalky ware shards have been classified into two categories according to their colour. In this way, chalky ware shards were either whitish or reddish.

At Site A, 160 whitish and 6 reddish shards were found in three strata (L3, L4 and L5) of TP A1. The chalky ware shards were determined as being pottery shards and basin shards with a foot ring. In terms of decoration, the vast majority of the whitish shards and all the reddish shards are plain. However, a small number of the whitish shards from the foot ring were decorated and perforated (see Figure 5.5). The incised design on these shards, which consists of concentric circles and arc pattern (see Figure 5.5) is typical of the Late Phase of the Middle Neolithic Age in Hong Kong and the Pearl River Delta.


*Figure 5.5 Incised Chalky Wares (Left)Chalky Basin Foot Ring Shards (Right) Identified in TPA1 at Site A*

Similar incised chalky and incised corded shards have previously been found elsewhere in Hong Kong such as at Fu Tei on Chek Lap Kok island ([50]) as well as at the Caotangwan (草堂灣) site of Sanzao (三灶) Island in Zhuihai City ([51]) in China. Using Carbon-14 dating techniques, the finds from Chek Lap Kok, were dated to between 3,355B.C. to 3,890B.C. and 3,650B.C. to 4,100B.C. ([52]). Both the original Fu Tei site, which no longer exists, and Site A of South Soko Island are situated at similar topography at a narrow terrace at an elevation of 12 mPD to 18 mPD facing the sea.

At Site G, two plain undecorated whitish chalky ware shards were unearthed. (see Figure 5.3(a), upper row). These shards were found in the interface between L3 and L4 of TP G2. Based on the presence of other pottery shards dated to the Bronze Age in this layer, the chalky shards were also determined as belonging to the Bronze Age.

Hard Chalky Shards

At Site E, a single piece of impressed raised square design as well as a single piece of net design pot shards were found at L3 of TP E1 (see Figure 5.6). The layer where these two shards were found also contained coarse shards that were decorated with a triangular geometric pattern. The raised square design is typical of Bronze Age pottery in Guangdong Province and in Hong Kong. Similar finds with same the patterns were previously unearthed in Hong Kong at Tung Wan Tsai on Ma Wan Island and Hai Dei Wan on Lantau Island([53]) .

Figure 5.6 Net Design Pot Shards (Left) Impressed Raised Square Design (Right) Found in L3 of TPE1 at Site E

The same kind of pot and gui (a kind of deep circular vessel with handles) with raised square design was previously found at Hengling shan (橫嶺山) cemetery in Baluo (博羅) in Guangdong Province. These were dated to Middle Zhou to Spring and Autumn Periods, or 2,400 to 3,500 years ago. Thereby, the hard chalky shards found in L3 of TP E1 are dated to this period ([54]).

Burial Urn with Two Lids

A burial urn with two layers of lids was found as an archaeological feature, L2/F1 in TP B1 in Site B (see Figure 5.7). The height of the pot is 47 cm, the pot exterior surface is covered with brown cracked glaze and the style of such glaze is the same as the those produced from Qishi (奇石) kilns in Guangzhou dated to the Song Dynasty. The same kinds of finds were previously found in Sha Ha Site in Sai Kung in 2002 ([55]). This kind of pottery is seldom recorded in published reports.

Figure 5.7 Burial Urn with the Inner Layer Lid (Left) Drawing of Burial Urn and the Outer Layer Lid (Right)

However, according to personal communication with a Guangdong pottery expert, Mr. Liu Chengji, a former Guest Research Fellow of AMO, based on the shape of the pottery, this kind of burial pot was commonly produced in the 18^th Century in Guangdong. Thus, the burial pot and feature is considered to date back to the 17^th to 18^thcenturies during the late Ming to middle Qing Dynasties.

5.1.3 Porcelain

Celadon

Three celadon bowl rims (see Figure 5.8) were found in TP D1 of Site D. Based on the glaze, it is classified as Qingbin (青白) ware, which was produced in Fujian coastal kilns during the Song Dynasty (A.D. 960-1279)([56]).

In addition, a piece of light green crackled glazed bowl shard was discovered in Layer 3 of TP G2 at Site G, which was dated to the Song Dynasty (see Figure 5.3(b)).

Figure 5.8 Celadon Shards Identified in L2 of D1

Blue-and-White Porcelain

Blue-and white porcelain shards were found in TP C6 of Site C, D1 of Site D and E1 of Site E (see Figure 5.9). By comparing these finds to the same kind of finds in Hong Kong, it was ascertained that these shards date back to the late 19^th to 20^th centuries.

Figure 5.9 Blue and white Porcelain Shards Identified in TPC6 of Site C

5.1.4 Stone Tools

The stone tools found in this survey were pebble tools, natural tools, striking tools and polishing tools. The pebble tools are handheld-sized pebbles used as tools. They include chopping tools, hammers, pounders and grinders. Naturally occurring stones which were ready for use and made of sandstone were chosen to be whetstones and saddle-quern stone. Pebble stones were used for a long period of time at coastal and riverside areas throughout the Neolithic Age to Early Iron Age.

Chopping Tool

A stone chopping tool was unearthed at the interface between L3 and L4 of TP G2 at Site G. The chopping tool is a heavy core tool and was presumably used for wood chopping. This chopping tool is made of whitish coarse tuff. It is bifacially trimmed, with flakes removed from both dorsal and ventral surfaces to form a sharp edge. It is 124 mm long with a height of 98mm. The angle of its chopping edge is between 60 degrees and 70 degrees. A striking platform can also be discerned (see Figure 5.10).

Pebble Tools

Most of the pebble tools were found in TPs A1, A2 and A3 at Site A. There are 3 grinders found in TPs A1, A2 and A3 and 3 pebble hammers found in TPs A1 and A3 (see Figure 5.11).

Figure 5.11 Pebble Pounders in TPA1 at Site A

Whetstone

Two whetstones were unearthed in the test pits at South Soko. One was in TP A1 and one was in TP B5 (see Figure 5.12). The purpose of a whetstone is to polish the edges of tools. Whetstones are usually made of fine sandstone.

Figure 5.12 Whetstone Found in L4 of TP B5 at Site B

Stone Saddle-quern

A large stone saddle-quern made of sand stone was found on the slope of Site B (see Figure 5.13), the length of the stone is 90cm. It would have been used for food processing or for stone tool polishing.


*Figure 5.13* **Photography and Drawing of Stone Saddle-quern**

Scraper

A scraper (see Figures 5.14 and 5.15) made of dark grey tuff was discovered in L3 of TP E1. The scraper is made of a stone flake. A notch can be identified on the left side of the flake axis. The length of the scraper is 5.1 cm and its thickness is 0.8 cm. Some small, dark grey tuff flakes were also found associated with the scraper.

Figure 5.14 Scraper (Left) and Awl (Right)

Figure 5.15 Drawings of Scraper (Left) and Awl (Right)

Awl

An awl made of dark grey tuff was found in L3 of TP E1. This artefact had been made of a blade with a point and edge (see Figures 5.14 and 5.15). A ridge can be identified from the dorsal surface which is formed by a special technique to control the flaking.

The length of the awl (measured along the flake axis) is 5.1cm, the width is 2.1cm and the thickness is 1.5cm (measured form diffuse bulb to height point of the ridge of the dorsal surface).

Flake-knife

A flake-knife made of an ovoid shaped flake was found in L5 of TP D1 at Site D. (see Figure 5.16). The rock used to make the knife is dark grey tuff. The working edge and striking platform can still be identified. No traces of trimming can be observed but a few scars remained on both the ventral and dorsal surfaces. The length of this knife is 3.5 cm (measured along flake axis) and the width is 2.8 cm.

Figure 5.16 Flake-knife Identified in L5 of TP D1

Stone Adze

A broken stone adze edge (see Figure 5.17) was found in L2 of TP A3 at Site A. Both sides of the adze surface were smoothly polished and wearing remains can still be observed from the edges. The adze was made of a dark green sedimentary stone.

Figure 5.17 Stone Adze Identified in L2 of A3

5.1.5 Stone Ornament

Quartz Ring

Two broken quartz rings (see Figure 5.18) were found at Site B. One was in TPs B5 and one in B8. In TP B5, the ring was found in F2, close to a quartz disc, a quartz core and a few quartz flakes. According to the section feature of the rings, these rings belong to the first phase of ring production in Pearl River Delta coastal areas([57]) dated to the Late Neolithic Age. The same kind of artefact had been revealed on the same slope in 1983 ([58]).

Figure 5.18 Stone Core, Stone Disc, Broken Stone Rings (from Left to Right), Reflecting the Process of Ring Manufacturing

5.1.6 Stone Ornament Raw Material and Production Waste

Quartz Flake and Quartz Core

A few quartz flakes were found in L4 of TP B5. These were associated with a broken quartz ring and quartz discs. It is believed that these flakes were waste material from quartz ring production. A quartz core was found in the same stratum of the same pit. The core is a raw material for quartz ring production (see Figure 5.19).

Quartz Disc

There were 12 quartz discs found in Site B. 11 discs were unearthed from L4 of TP B5 and another one was identified from fieldwalking on a slope some 13 m to the southwest of TP B5.

In TP B5, 10 quartz discs were orderly placed next to a boulder. This indicated that the cluster of disc is an undisturbed in situ deposit (see Figure 4.7). These discs are waste material from quartz ring manufacturing that can be reused to produce more new quartz rings ([59]). Thus, the discs were deliberately placed next to the boulder by ancient people for storage to be collected and reused later. The diameter of these discs is in the range of 3.1 cm to 5.0 cm, the thickness is in a range of 1.2 cm to 1.4 cm.

Figure 5.19 Quartz Discs Identified in F2/L4 of TP B5

Another quartz disc was picked up in a terrace field cut section (see Figure 4.5, right) and it is believed that it was washed down from the upper slope.

By comparing the finds to those found at the Yung Long Site at Tuen Mun and Baojianwan in Zhuhai City ([60]), it was determined that these finds date back to the Late Phase of the Late Neolithic Age.

Since artefacts from different stages of quartz ring production were found, it was considered that a quartz ring workshop site occurred on South Soko Island dating back to the Late Neolithic Age (some 4,000 years ago).

5.1.7 Shells

A total of 1,803 shells were found in L3 of TP E1 at Site E. They are classified into 14 species by visual examination. Based on a hard raised square pattern pot shard and some triangle pattern pot shards associated with the shells, these shells are dated to Late Bronze Age (2,500 to 3,000 years ago).

Some of the shells have burn marks which reflected that they were collected for food consumption. Table 5.2 summarises the shell species identified and some representative specimens are shown in Figure 5.20.

Table 5.3 Shell Species and Quantity

Species	Quantity	Habitation Area
Nerita albicilla	824 (45.7%)	Rocky shore, lagoon
Monodonta labio (Common Top Shell)	279(15.47%)	Rocky, cobble and pebble shore
Cellana toreuma(Common Limpet)	192(10.64%)	Rocky shore
Chlorostoma argyrostoma (Orange Top Shell)	159 (8.81%)	Rocky and cobble, pebble shore
Thais luteostoma (Yellow-mouthed Whelk)	104 (5.76%)	Rocky shore
Camaena cicatoricosa (Large Striped Snail)	98 (5.43%)	Forest
Nerita costata	45 (2.48%)	Rocky shore
Lunella coronata (Common Turban Shell)	40 (2.21%)	Pebble shore and sandy shore
Collisella sp. (Limpet)	19 (1.05%)	Rocky shore
Collisella dorsuosa (Limpet)	19 (1.05%)	Rocky shore
Chicoreus torrefactus (Common Murex)	4 (2.2%)	Rocky shore
Cronia margariticola ( Small Blue Mouthed Whelk)	2 (1.1%)
Strombus Luhanus (Blood-mouthed Stromb)	2 (1.1%)	Sandy and muddy shore
Polynices melanostoma (Blotched Moonshell)	1 (0.55%)	Sandy shore
Unclassified	15 (0.83%)
	Total 1,803





*Figure 5.20 Shells Identified from L3 of TP E1 at Site E*

5.1.8 Fish Bones

There were 206 pieces of crania and two jawbones of Pomadasys hasta (head grunt) identified in L3 of TP E1 at Site E. (see Figure 5.21). During spring and early summer, P. hasta migrate shoreward to spawn, by which time they would have been readily accessible in the shallow inshore water. Their spawning season corresponds with the beginning and the initial peak period of the wet southwest monsoon (June-July) ([61]).


*Figure 5.21 Head Grunt Head Bones and Jaw Bone in L3 TP E1 at Site E*

5.1.9 Animal Bone

There were 15 pieces of animal bones discovered from L3 of TP E1 and the bones had been burned (see Figure 5.22). This indicates that area of E1 was used as rubbish dump. According to the datable pottery associated with the bones, they date back to Bronze Age, some 3,000 years ago.

Figure 5.22 Animal Bone in L3 of TP E1

5.1.10 Animal Tooth

An animal tooth (see Figure 5.23) were found in L3 of TP E1. However, the species of this animal could not be determined.

Figure 5.23 Animal Tooth in L3 of TP E1

5.1.11 Archaeological Features

An archaeological feature refers to the remains of a structure or burial ground. A total of three features were found in this survey at Site B, one from TP B1 and two from TP B5.

One of the features, recorded as L2/F1 of TP B1, is a secondary burial. This burial practice was commonly adopted in the Guangdong Province for hundreds of years. The feature is a circular pit where a burial pot was found (see Figure 4.8). The height of the pot is 47 cm with brown cracked glaze on the exterior (see Figure 5.7). The glaze of this pot is similar to the Qishi ware produced in Guangzhou during the Song Dynasty, which was also discovered in Sha Ha Site in Sai Kung in 2002([62]). However, due to the fact that the shape of the pottery was the same as those commonly produced in the Guangdong region in the late 17^th to 18^th centuries, it is dated to late Ming to Middle Qing Dynasties.

The other two features were found in L4 of TP B5 in Site B. These were recorded as L4/F1 of TP B5 and L4/F2of TP B5. L4/F1 contains a cluster of pebbles, with a few quartz flakes, a quartz disc, a quartz core and two pieces of pebble refitting flakes. A broken quartz ring was also found near to the feature. The function of this feature is uncertain. However, it reflects the different stages of quartz ring manufacturing process (see Figures 4.7 and 5.18). L4/F2 of B5 contains a large boulder, a few rocks and a cluster of 10 quartz discs beneath the large boulder (see Figures 4.7 and 5.19). Based on the typology of the quartz ring and discs, F1 and F2 date back to the Late Phase of Late Neolithic Age.

5.2 Shek Pik

The representative special finds identified in Shek Pik are described below.

5.2.1 Pottery

Glazed Ware Shards

There were four pieces of glazed pot dark brown and reddish colour shards found in L2 of SP1 (see Figure 5.24). The glaze and shard texture indicates these pieces date back to the late 19^th century to modern period.

Figure 5.24 Dark Brown Glazed Ware Shard (Left, 3 pieces pieced together)

Coarse Ware Shards

There were three pieces of coarse ware shard found in L2 of TP SP4 (see Figure 5.25). No decoration pattern can been found on the shards surface. The shard section is black in colour and coarse sand was observed in the section. These shards are comparable with the same kind of finds found in Hong Kong and the Pearl River Delta archaeological sites dated to the Late Neolithic Age.

Figure 5.25 Coarse Ware Shards in L2 of SP4 at Shek Pik Tung Wan

5.2.2 Lime Kiln Structure Remains

A total of three broken kiln bricks and a broken fired bar were found in L2 of TP SP1 (see Figure 4.21). These finds indicate the presence of a kiln at the site which dates back to Tang Dynasty (A.D. 618-907). Such an industry was very common in coastal areas in Hong Kong at that time.

5.2.3 Stone Tools

Pebble Grinder

A grinder stone was found in L4 of TP SP1. The length of the stone is 9.7 cm. It was struck and a diffuse bulb of force is observed. Owing to the flat surface, the pebble stone is identifiable as a grinder (see Figure 5.26).

Figure 5.26 Pebble Grinder in L3 of SP1 at Shek Pik Tung Wan

5.2.4 Flakes

Three stone flakes (see Figures 5.27 to 5.29) were found in L4 and L5 of TP SP1. The first one was struck from a pebble. Its length is 13cm with a natural surface. A cutting edge was identified which resulted from flaking. A scar can be seen on the flake.

The second flake is made of light green tuff. Its length, measured along the flake axis, is 7.3cm. The thickness is measured at the highest point of force bulb, which is 1 cm. A striking platform, a striking point and a bulb of force are found on the flake. A working edge with four notches is observed at the left side of the flake axis.

The third flake was found from L5 of TP SP1. It was made of light green tuff. The length of the flake is 12.1cm and the thickness is 3.8cm. A striking platform and a natural surface are found on the flake. A flaking scar can be seen on both the dorsal and ventral surfaces. A ridge exists on the dorsal surface which is regarded as a feature of the flaking control technique. The function of the flake is uncertain.

Figure 5.27 Stone Flakes in L5(Left) and L4 (right) of SP1 at Shek Pik Tung Wan

Figure 5.28 Drawings of Stone Flakes in L5(Left) and L4 (right)

Figure 5.29 Stone Flake from L5 of SP1 at Shek Pik Tung Wan

6 ConclUsions

A total of 55 auger holes and 61 test pits and, 19 auger holes and 7 test pits were undertaken on the South Soko Island and Shek Pik respectively for the archaeological surveys between 2004 and 2006. Based on the survey findings archaeological deposits were identified in all the sub-sites on South Soko Island (i.e. Sites A to G) as well as the Shek Pik area within the Shek Pik Tung Wan Archaeological Site (Site F). The archaeological extent identified in each area is presented in Figures 4.4 and 4.27.

On South Soko Island evidence from the archaeological features, artefacts and ecofacts identified indicate that human activities on the island as early as 5,000 years ago. Discoveries of archaeological features, artefacts and ecofacts are not only from one period of time but span a range of periods including the Middle Neolithic Age (4,000B.C. – 2,500B.C.), Late Neolithic Age (2,500B.C. -1,500B.C.), Bronze Age (1,500B.C.-221B.C.), Song Dynasty (A.D.960-1279) and late Ming to Middle Qing Dynasties (the 17^th to 18^th centuries).

Pottery artefacts identified in South Soko Island comprised corded and geometric coarse ware shards, incised chalky pot or basin shards, celadon shards, porcelain shards and a burial urn. A variety of stone artefacts were identified which include a chopping tool, pebble pounders, pebble hammers, stone flakes, whetstones, grinding stones, stone saddle-quern, scraper, flake-knife, broken blades, stone quartz rings, quartz discs and a quartz core. Ecofacts identified include shells, fish bones and animal bones and tooth.

At Shek Pik, archaeological potential exists at the raised beach area and the foothill area within the Shek Pik Tung Wan Archaeological Site. Artefacts identified include lime kiln remains dated to Tang Dynasty (A.D. 618-907) and coarse ware shards and stone flakes dated to the Late Neolithic Age. However, no archaeological potential is revealed from the survey at the area adjacent to the Declared Monument, the Shek Pik Rock Carving.

([1]) Cross-dating refers to “Artefacts from an archaeological site are often dated by correlation with typologies of similar artefacts in the surrounding area. This method is based on the assumption that typologies evolved at the same area”. From Whitehouse, R.D. ed 1983 Macmillan Dictionary of Archaeology, London: Macmillan Press.

([2]) Schofield, W. 1925-38 Transcript of Handwritten Notes on Hong Kong Archaeology. (Unpublished)

([3]) Schofield, W. 1940 The Proto-Historic Site of The Hong Kong Culture at Shek Pek, Lantau, Hong Kong. Singapore: Proceeding of the Third Congress of Pre-historians of the Far East.

([4]) Peacock, B.A.V. and T.J.P. Nixon 1995, The Hong Kong Archaeological Survey: Subsurface Investigation Reports, Hong Kong: AMO.

([5]) Peacock, B.A.V. and T.J.P. Nixon 1995, Ibid.

([7]) Peacock B.A.V. and T.J.P. Nixon 1995, Ibid.

([8]) Peacock, B.A.V. and T.J.P. Nixon 1995, Ibid.

([9]) 廣東省文物考古研究所 1997 《香港南丫島、長洲島、大鴉洲島、小鴉洲島及石鼓洲島考古調查報告》，香港，古物古蹟辦事處 (未刊)。

([10]) 舒懋官編纂 1819 《新安縣志》，新安縣衙。

([11]) Lockhart, S. 1898 Report on Extension of The Colony of Hong Kong, London, UK Colonial Office. (unpublished)

([12]) Empson, H 1992 Mapping Hong Kong: A Historical Atlas, Hong Kong, Government Printer.

([14]) Wilkinson, Julia 1989 “The Ghost Isles”, Another Hong Kong – An Explorer’s Guide. Hong Kong: Emphasis (HK) Ltd.

([15]) Peacock, B.A.V. and T.J.P. Nixon 1995, The Hong Kong Archaeological Survey: Subsurface Investigation Reports, Hong Kong, Antiquities & Monuments Office, 67-82.

([19]) Langford, R.L. and others 1995 Geology of Lantau District, Hong Kong, Geotechnical Engineering Office.

([20]) W. Schofield 1940 The Proto-Historic Site of the Hong Kong Culture at Shek Pek, Lantau, Proceedings of the Third Congress of Prehistorians of the Far East, Singapore.

([21]) Chen Kung-chieh 1957 Archaeological Surveys and Excavations in Hong Kong, Kaoku Hsueh Poa, No.4, 1957.

([22]) Frost, R.J. 1980 “Shek Pik”, Journal of Hong Kong Archaeological Society, vol.VIII, Au Ka Fat and others 1989 “Neolithic Site in Tung Wan”, Journal of Hong Kong Archaeological Society , vol. XII. ; 區家發、鄧聰 1988 〈香港大嶼山東灣新石器時代沙丘遺址的發掘〉《紀念馬壩人化石發現三十週年》，北京，文物出版社。鄧聰 1991 《香港考古之退旅》，香港，區域市政局。

([23]) Chau Hing Wah 1993 “Periodization of prehistoric Culture of Pearl River Delta Area”, Collected Essays on the Culture of the Ancient Yue People in South China, Hong Kong, Urban Council.

([24]) Dreweet, Peter and others 1997 South Lantau Archaeological Survey 1997 (Unpublished)

([25]) 郭棐 1573-1620 、1997《粵大記》卷32〈政事類‧海防卷〉，廣州，中山大學出版社。

([26]) 蕭國健 1985 《清初遷海前後的香港之社會變遷》，台北，台灣商務印書館。

([27]) 靳文謨編纂 1688 《新安縣志》，新安縣衙。舒懋官編纂 1819 《新安縣志》，新安縣衙。

([28]) Lockhart, S. 1898 Report on Extension of The Colony of Hong Kong, London, UK Colonial Office. (unpublished)

([29]) 蕭國健 1985 《清初遷海前後的香港之社會變遷》，台北，台灣商務印書館。

([30]) Peacock and Nixion 1985 Report of the Hong Kong Archaeological Survey, Vol. III, Part 3, Hong Kong, AMO.

([31]) 商志香覃、毛永天 1997 〈香港地區新石器文化分期及與珠江三角洲地帶的關係〉，《考古學報》，1997年3期。

([32]) Peacock and Nixion 1988, 1995 The Hong Kong Archaeological Survey: Subsurface Investigation Reports, Hong Kong, AMO.

([33]) Peacock and Nixion 1988, 1995, Ibid.

([34]) 鄧聰 1992 《香港考古之旅》，香港區域市政局。

([35]) 商志香覃毛永天 1997 〈香港地區新石器文化分期及與珠江三角洲地帶的關係〉，《考古學報》，1997年3期。

([36]) Meacheam, W. ed 1978 Sham Wan, Lamma Island: An archaeological Site Study, Hong Kong Archaeological Society.

([37]) 商志香覃毛永天 1997 〈香港地區新石器文化分期及與珠江三角洲地帶的關係〉，《考古學報》，1997年3期。

商志香覃 2000〈香港地區青銅時代考古及有關問題的探討〉，《香港考古論集》，北京，文物出版社。

([38]) 廣東省文物考古研究所 1997 《香港南丫島、長洲島、大鴉洲島、小鴉洲島及石鼓洲島考古調查報告》，香港，古物古蹟辦事處 (未刊)。

([39]) Meacham ,W ed 1978 Sham Wan, Lamma Island, An Archaeological Site Study, Hong Kong Archaeological Society.

([40]) Au Ka Fat and others 1990 “Neolithic Site in Tung Wan, Shek Pik, Hong Kong, Brief Report of the 1987, 1988 Excavations”, Journal of Hong Kong Archaeological Society, Vol.XII.

([41]) 鄧聰 1992 《香港考古之旅》，香港，區域市政局。

([42]) Frost, R.J. 1979 “Tung Wan(Shek Pik)”, Journal of Hong Kong Archaeological Society, Vol.VIII.

([43]) 鄒興華 1993 〈珠江三角洲史前文化分期〉，《嶺南古越族文化論文集》，香港，市政局。

([44]) Meacham W. ed 1979 Sham Wan , Lamma Island, An Archaeological Site Study, Hong Kong Archaeological Society.

([45]) Meacham, W and others 1994 Archaeological Investigation on Check Lap Kok Island, Hong Kong Archaeological Society

([46]) Rogers, P.R. and others 1995 Tung Wan Tsai : A Bronze Age and Han Period Coastal Site, Hong Kong, Antiquities and Monument Office.

([47]) Meacham W. ed 1979 Sham Wan , Lamma Island, An Archaeological Site Study, Hong Kong Archaeological Society.

([48]) Meacham, W and others 1994 Archaeological Investigation on Check Lap Kok Island, Hong Kong Archaeological Society, Rogers, R.R. and others 1995 Tung Wan Tsai: A Bronze Age and Han Period Coastal Site, Hong Kong, Antiquities and Monuments Office.

([49]) 朱非素 2003 〈試論石峽遺址與珠江三角洲古文化的關係〉，《廣東省文物考古研究所建所十週年文集》，廣州，嶺南美術出版社。

([50]) Meacham, W 1994 Archaeological Investigations on Chek Lap Kok Island, Hong Kong Archaeological Society

([51]) 珠海市博物館等 1991 《珠海考古發現與研究》，廣州，廣東人民出版社。蕭一亭 2004 《先秦時期的南海島民---海灣沙丘遺址研究》，北京，文物出版社。

([52]) Chau Hing Wah 1993 Periodization of prehistoric Culture of Pearl River Delta Area, Collected Essays on the Culture of the Ancient Yue People in South China, Hong Kong , Urban Concil.

([53]) Williams, B. 1980 “Hei Dei Wan” , Journal of The Hong Kong Archaeological Society, vol. VIII.

([54]) 廣東省文物考古研究所 2005 《博羅橫嶺山-----商周時期董地2000年發掘報告》，北京，科學出版社。

([55]) 廣東省文物考古研究所 2005 〈西貢沙下遺址出土宋明陶瓷的考古學觀察〉，《香港的遠古文化---西貢沙下考古發

現》，香港，古物古蹟辦事處。

([56]) 香港城市大學中國文化中心 2004 《陶瓷下西洋》，香港城市理工大學。

(1) Peacock and Nixon 1986 " Tai A Chau", The Hong Kong Archaeological Survey: Subsurface Investigation Reports, Hong Kong, Antiquities & Monuments Office.

([58]) Peacock and Nixon 1986 " Tai A Chau", The Hong Kong Archaeological Survey: Subsurface Investigation Reports, Hong Kong, Antiquities & Monuments Office.

([59]) 鄧聰 2000 〈環珠江口考古之崛起，玉石飾物作坊研究舉隅〉，《珠海文物集萃》，香港中文大學。

([60]) AMO 1997 “Excavation at Neolithic Yunglong Site, Hong Kong”, Kaogu (Archaeology) No.6, 1997.

廣東省文物考古研究所珠海市博物館 2004 《珠海竇鏡灣---海島型史前文化遺址發掘報告》，北京，科學出版社。

([61]) Chan, William 1978 Identification and Analysis of Excavated Fish Remains, Sham Wan , Lamma Island, An Archaeological Site Study, Hong Kong Archaeological Society.

([62]) 廣東省文物考古研究所 2005 〈西貢沙下遺址出土宋明陶瓷的考古學觀察〉，《香港的遠古文化---西貢沙下考古發現》，香港，古物古蹟辦事處。