- Email: [email protected]
Stone anchors near Black Fort, Galle, Sri Lanka
The hTternationalJournal of Nautical Archaeology (1998) 27.4:331-342 Article No. na990187
Stone anchors near Black Fort, Galle,. Sri Lanka
®
Corioli Souter Australian National Centre of Excellence for Maritime Archaeology, Western Australian Marithne Museum, Cliff Street, Fremantle WA 6160, Australia
Introduction The Joint Australia-Sri LankaNetherlands Maritime Archaeological Project is centred in the World-Heritagelisted Galle Harbour, Sri Lanka. The harbour is notable for the VOC (Dutch East India Company) shipwrecks it contains and has been the subject of underwater archaeological surveys since 1992. Galle was an operational port and major entrepot in pre-Christian times with the area coming to prominence in the 12th century with the arrival of the Portuguese and later the Dutch. Galle was the second most important harbour of the VOC, strategically located in terms of trade between the Arabian Peninsula and East Asia. Galle remained the major port in Sri Lanka after the arrival of the British in 1796 and it was only in 1873 when the present port of Colombo was constructed that Galle became marginalised. During the 1996 season, a side scan sonar survey (Green, 1998) led to the discovery of a stone anchor site. A S C U B A circular search to investigate a side scan anomaly resulted in the discovery of one large Arab-Indian stone anchor, G H P 30 (Fig. 1), and a broken portion from a smaller anchor of the same type. The Arab-Indian anchor type has been dated to the period from the 6 t h t o 16th century A D (Vosmer, 1998, pers. comm.). ,The larger anchor was recorded as 350 cm long, the broken anchor approximately 150 cm. Both were located close to the Black Fort, 1057-2414/98/040331 + 12 $30.0010
in the area of K a m b a Bandina Gala, a traditional mooring site for vessels lightering cargo to and from Galle. In the same general area is a site known to have ceramics from a wide geographical and temporal range. This could possibly be associated with the mooring of vessels and the stone anchors. When discovered, the anchors lay flat and half-buried on a sand seabed covered with a layer of fine silt and organic material. During the 1997 season it was planned to raise these stone anchors and to determine whether there were any inscriptions; for petrological studies to ascertain the origin of the stone; and conservation. Following their conservation it was intended that the anchors be incorporated into a display for the Maritime Museum in the Galle Fort.
Site description The Galle Harbour seabed consists of a mixture of sand, sediment and rock overlying mud deposits. A small limestone reef runs through the middle of the anchor site. At the time of survey the underwater visibility was extremely poor, ranging between 0 m to 1"5 m at best. The visibility improved in the week prior to full moon, indicating that it was influenced by tidal and subsequent sediment movement. Sri Lankan Sub Aqua Club divers, having made underwater observations in the harbour following the monsoon season, reported considerable upheaval of the seabed resulting in exposure of shipwrecks © 1998 The Nautical Archaeology
Society
N A U T I C A L
A R C H A E O L O G Y ,
~----;< ~;
"
. i..:'.:..,~ ':'i
27.4
• ."::~'
.'l
...
I
•
~
"
'~
L
.. ?'.
.
..
"- -": "-;L: '
I
-,-..'-!-.:~.....-?:~-:,T:.;.:...::.r..,.,,,,.~.x,..~:~.....,~,~.~....~.. • • , , . . . '
; !!!:;v:: : • . .
"
" .';*~;
.
.
" "•
.
'
". '
..r,
.
.... ". ""."." "
"
•
..
. . . .
' ' : " i!i':
I
I
0
50
" • "-~:'~,."', r " ; ~ ~ ' "'
•
,~....:..~.-~ ,.. :
": . . . .
•
. . . .
...~
"
•
..~
.
' ~
- . - . : , : = . . . ~
......:i::i:"i
.....
c m
Fig~tre 1. Line drawing of GHP 30.
and dispersal of artefacts. The anchor site is approximately 7 m deep and is interspersed with broken ceramics, stone fragments, iron concretions, firebox slag, bones, coal and modern rubbish. Although the area is subject to surge, the seabed is generally level except where the presence of large objects, such as the stone anchors, has resulted in scour or build-up of sediment on one face. Search
and
survey
During the 1997 season, diver search methods were employed to relocate the anchors. Circular jackstay searches and swim lines were used to ensure systematic coverage of the seabed in low visibility, to locate the objects. The large stone anchor GHP 30 was relocated first. Another stone object, approximately 50cm in diameter, was found 14 m from G H P 30. The round hole 332
in this stone suggests that it might be another broken portion of an anchor. The object was identified as a modern mooring upon recovery, with a small, machine drilled, round hole and a modern piece of chain concreted to the surface of the rock. During this search, a third previously unrecorded stone anchor, G H P 32, was found. It was a flat trapezoidal anchor with two round vertically- aligned holes at the top and two square horizontallyaligned holes located in the lower half (Fig. 2). While the two round holes present an unusual configuration, the shape is typical of the Mediterranean. The search continued m an easterly direction from this anchor in an attempt to locate the broken anchor found in 1996 (Fig. 3). A crown of an Arab-Indian anchor which had broken off at the second square hole from the top was discovered.
C. SOUTER: STONE ANCHORS
Exposed side
Underside
I
50
0 cln
Figure 2. Line drawing of G H P 32.
I ?
, I
' I I
I I
', [
) Profile
I
I
0
I
50
cm Figure 4. Line drawing of G H P 31. 0
5O
Figure 3. Line drawing of 1996 broken anchor in situ.
However, the dimensions of this anchor, 74 cm x 37 cm, were smaller than those of the broken anchor recorded during the 1996 expedition This suggests that this was a new find and not the original example being sought. The difficulty in identification was naturally compounded by the poor visibility. This new anchor was recorded and labelled G H P 31 (Fig. 4).
During searches south of the large anchor G H P 3 0 ' s position, a round anchor, GHP 76, with a diameter of 65 cm and height of 15 cm was also discovered. G H P 76 is essentially a rounded stone with a squared off hole through the centre (Fig. 5). An 18th-century admiralty iron anchor was also found in this area. The relative positions for each anchor, the modern mooring and the rock outcropping from the seabed were recorded by trilateration. From these anchor positions subsequent circular searches could be used 333
NAUTICAL ARCHAEOLOGY, 27.4
I 0
cm
I 50
Figure 5. Line drawing of G H P 76.
to locate further artefacts and to better define the site by location and mapping of the natural features. All positions were incorporated into the Rule Web program and the resultant co-ordinates plotted into a site plan (Fig. 6) Anchor recovery
In preparation for raising, rope was tied around the largest anchor, G H P 30, at three intervals along the length of the shank. A space under the anchor G H P 30 was cleared using a water jet in order to attach the ropes. During this procedure a substantial piece of blackened timber was located. It was thought that this wood could be a remnant of the anchor arm or stock. Further excavation of the area under G H P 30 revealed two pieces of hardwood. Hardwood timber would be strong and durable enough for anchoring purposes. The larger piece GHP 45 has been identified as belonging to the Leguminosae family, best matching the Xylia species (pyinkado), common in tropical Asia. The species is known as irul in India which suggests its presence on the sub-continent and possibly Sri Lanka. The smaller timber, GHP44, was also identified as belonging to the Leguminosae family and is most likely to be Sindora species (sepitir, sindoer). This species is native to South East Asia. There is no reference to the species being indigenous to Sri Lanka or 334
India (Bolza & Keating, 1982). This is the first time wood has been found in association with an Arab-Indian stone anchor. It is hoped that the discovery of timber will provide insight into the configuration of these anchors. The timber's position in relation to the stone shank and in particular the square holes suggests that the wood functions as the arms of the anchor. The first piece raised, G H P 4 4 (Fig. 7), is 980 mm in length, 160 mm at its widest point and 104ram high. It was found approximately 300mm under the sediment. There is no uniform stratigraphy as the first layers of sediment contain a random mix of broken ceramics and roofing tiles. The lack of stratigraphy implies that the area is, or has been, subject to a high degree of disturbance. The uppermost surface of the timber as found & situ exhibits curvature with most of the original surface intact. The surface also exhibits tool marks, probably those associated with adzing. The underside has been partially degraded by marine borers although much of the original surface remains. The original shape of the timber has been retained and has a natural concave curve on one edge approximately 73 cm along its length. It appears as though the curve of the timber has been tooled and the edges bevelled to shape an arm/stock. It is difficult to ascertain the original shape at this end of the timber as a large section is missing. However, in profile view the end appears originally to have been finished in a chisel point. With this in mind, the missing portion of timber may have been symmetrical with that which remains, together forming a long, gradually tapered, V-shaped anchor arm or possibly stock. The thicker end of the timber, which was found next to the shank, is mortised and bears extensive degradation by marine borers. It appears that the mortised end has snapped off suggesting that this piece is only a portion of the original timber. The function of the mortised end is unclear. It
C. SOUTER: STONE ANCHORS
Modern
Mooring @
'IN
GI-IP 30
GHP32
* GHP31 Broken anchor in situ
:~
~Iron
anchor in situ
GHP 76
I
50
lIl
Figure 6. Site plan.
0
50 cm
Figure 7. Line drawing of GHP 44.
may have performed as a locking mechanism to secure the timber into the anchor shank. The second piece raised, G H P 45, is considerably larger than the first (Fig 8). It is 134 cm in length, 26 cm wide and up to 18 cm high. The timber is similar to the smaller piece as it exhibits a natural curve, which has been enhanced by tooling. The
I
1
0
50
cm
Figure 8. Line drawing of GHP 45.
timber has retained its original surface, in part. Raised knots, up to 5 m m high on the upper surface of the wood as found in situ, may indicate where the original surface was and the subsequent extent of degradation, although it is also conceivable that they are typical wood growth formations and part of the timber surface. The grain of the timber is visible on the surface of the 335
NAUTICAL ARCHAEOLOGY, 27.4
thicker end. This end lay perpendicular to the anchor, under the stone shank just behind the crown. It is possible that either it has broken off from another piece of timber or it has broken off from the hole in the anchor shank, as it was excavated from beneath the anchor. The curve of the timber begins 65 cm from the wider end to form what may be interpreted as an arm. The timber may have been selected for its naturally contoured shape and then mechanically worked to create an arm, which would penetrate and grip into the sediment. The end appears to finish in a point, although this, too, is broken or worn. Tool marks are visible particularly in profile, although not to the same extent as the smaller timber. The underside of the large timber, as found in situ, shows evidence of degradation due to marine borers with some of the original surface remaining. It curves upward contributing to the hooked-arm appearance of the timber. Timber samples were collected for identification, CI4 dating and dendrochronological analysis. The radiocarbon age determination for the sample from GHP44 was 4 3 0 + 8 0 B P , by The University of Waikato Radiocarbon Dating Laboratory (1998). The timbers as we find them are incomplete, which prevents overall size determination. Both pieces are too small to fit well within the square holes (Fig. 9). It is also likely that the timber has swelled since immersion. If this were the case the arms may have been secured in the holes by a wedge or tenons as suggested by the mortised end of G H P 44.
Stone anchor descriptions G H P 3 0 is made from a sedimentary polymict conglomerate and is 317 cm in length. It is 55 cm at the crown and tapers to 26 cm at the opposite end. G H P 31 is composed of a sedimentary sandstone beachrock and has much smaller dimensions being 74 cm long. It is 30 cm wide at 336
Figure 9. GHP 30 with timbers. (Photo: Patrick Baker)
the hawser end and a maximum of 37 cm at the break, approximately half-way up the shank. The anchors are oblong in shape with a round hole in one end, presumably to take the mooring rope, and two square holes in the other end set at 90 ° to each other, one slightly further up the shank than the other. Through these opposed facing holes wooden stakes were set to act as the arms or the stock of the anchor. The type was initially identified by Honor Frost as a grapnel due to the transverse square holes in the lower part of the shank. This has been disputed as these square holes are different sizes and the anchor shanks are usually rectangular in cross section while the shafts of fourarmed grapnels are square or round in cross section (Kapit/in, 1994: 2). They are also somewhat smaller in length. Kapitfin has proposed a reconstruction where the larger hole near the end of the shank actually holds a stock rather than an arm (Kapit/in, 1994: 2). The discovery of wood in association with G H P 30 suggests a different reconstruction. The in situ positions of the two baulks of wood in relation to
C. SOUTER: STONE ANCHORS
the anchor holes indicate that the arm came from the hole nearest the crown. The position of the timbers on the seabed suggests that the lower hole on the stone shank held a curved wooden arm and the upper hole held a smaller piece of wood, which may be either an arm or stock. The round hole at the other end of the shank would have had a rope directly tied through it or perhaps another stock. The suggestion of a stock through this hole is evidenced by other anchor examples such as a shank found in the Red Sea at Lone Mushroom, west of Ras Muhammed (Raban, 1990: 302). This Lone Mushroom anchor is rectangular in cross section but only has one rectangular hole in the lower part that passes at right angles to the axis of the circular hole at the other end. This arrangement suggests that the rectangular hole held the arm of the anchor while the round hole held the stock. A timber would be required in the round hole in order for the anchor to grip into the seabed effectively (Kapit~in, 1994: 2). Alternatively, the notches at the base of the crown on the Lone Mushroom anchor may have been used to lash another timber to the base to be used as an arm/stock, or they may simply be quarry marks. However, the size of the circular holes in the Sri Lankan examples, suggests that only a rope would be passed through them. G H P 3 2 is a four-hole composite anchor, which essentially means that the anchor does not rely on its weight alone (Nibbi, 1991: 191). It is constructed from sedimentary sandstone and is 100 cm high. The maximum width is 83 cm tapering into 40 cm near the top. The anchor has a thickness of 17 cm. The type has been labelled Byzantine-Arab, as they were common in this period (Frost, 1963: 49). They were still in use until recently, both in the Persian Gulf and more commonly the Mediterranean, and as a result have also been called Mediterranean in form (Frost, 1963: 13).
G H P 32 has a pock-marked surface as a resuli~ of exposure and encrustation. Only a small portion of the original surface remains. The anchor has two round holes below the apex and two square holes along the base. It is generally accepted that the top hole(s) in the anchor, were for the main hawser. This is an unusual example as this style of anchor as found in other contexts has only one round hole below the apex. This may be a deliberate double-hole configuration to ensure the anchor stone is secured to the lifting rope, or perhaps the lower hole was added later to replace the original hole which is only 45 mm from the top of the anchor. If the original hole was used on its own it may have caused the stone to break under strain. The anchor has random channels cut into the surface, possibly wear marks from rubbing of the hawser rope or ropes on the stone surface while the anchor was tethered on the sea floor or to the ship during transportation. The channels may have resulted from lashing the timber arms to the anchor and subsequent wear by the rope. Alternatively, the channels may be the result of environmental conditions affecting the stone in situ such as sand built up on the face of the anchor scouring out the sandstone. The square and round holes also show signs of wear. G H P 76 is a round, stone anchor with a single, squared perforation. It is constructed from sedimentary sandstone with the dimensions 65 x 49 x 15 cm. This round anchor is possibly a weight anchor• This type relies on its weight as opposed to arms to hold a vessel. Its shape and size may indicate that it was used for a small vessel on a rocky sea floor (Vosmer, 1997: 9). The purpose of the hole's square finish is unclear• The shape of the hole is squared off perhaps so that a wooden arm or stock may be inserted and locked into place. In comparison, the holes in the Late Bronze Age Ugaritic anchors (Frost, 1970: 388) are cupular drilled yet chiselled square on 337
NAUTICAL ARCHAEOLOGY, 27.4
one side perhaps to hold the wood in place which was common to the Minoan Period and stop it loosening through rotation. (Nibbi, 1993: 14). I f G H P 76 is a fragment However, in the Ugaritic anchors this tech- of a larger anchor it is also conceivable nique was thought to be decorative as the that it may be the remains of an A r a b original drill holes are narrow and round Indian type. The size of the anchor in and not cut to fit square flukes (Frost, cross-section appears to negate this, being 1970: 388). There is no evidence to support marginally smaller than other known exthis in other single hole anchors. However, amples. It is not clear, however, how fragsquare holes in the Mediterranean-style mentary this example is. If the anchor is anchors have generally denoted the use of classed as Arab-Indian it would also better wooden arms. It is not necessary to square explain the square hole. In terms of function of an anchor, it is off a hole if it is to be used for a hawser alone. The'shape of the anchor has some the chain or rode as opposed to the anchor similarities to the ringstones used in the weight itself which holds a ship in position Maldives (Vosmer, 1997: 17). Ringstone (Green, 1971: 175). G H P 7 6 may have anchors were used for anchoring on coral been used in conjunction with several simireefs to obviate snagging on the coral. A lar one-hole anchor types to weigh down forked branch was inserted through the the anchor rode to allow it to hang parallel hole of a ringstone operating as a crude with the sea-bed. With this system, a cbmshank. (Vosmer, 1997: 17). The ringstone posite anchor would be attached to the end anchor comparison, however, fails to of the rode with a number of one-hole address where G H P 76 was found. The anchors attached to the rope at intervals sand and sediment seabed of Galle (Green, 1971: 175). This type of simple Harbour is unsuitable for either rock or single hole anchor is also in use today and coral anchor. Its function as found in situ is makes dating G H P 76 difficult. probably explained as a temporary mooring or abandoned anchor for a smaller Petrological analysis boat. It is also possible that it is not in its Most stone local to Sri Lanka is igneous. original form. It may be a fragment, such The stone sample from the large anchor G H P 30 has been identified as a sedimenas the lower corner of a Mediterranean three-hole anchor, similar to G H P 32. This tary conglomerate by Sri Lankan geolowould explain the square hole and the gist, A. Guntalika (Green, 1998: 35). This unusual shape. The closest comparison to type of stone is quite common in alluvial this anchor comes from Kommos in fan or fan delta complexes flanking rising Southern Crete and comes from a defin- mountain fronts or fault scarps. From able chronological context (Shaw, 1995: initial inspection the the most probable 282). It was dated to the late Minoan source of the anchor stone is Oman, as Period of 1250 BC (Nibbi, 1993: 13). The similar conglomerate formations occur Kommos example, artefact Number $636, along the mountain front running along is one of five anchor fragment finds and it the Omani coastline (Green, 1998: 35). A exhibits the same characteristics including geologist in Oman also examined the the squared off hole, as the Sri Lankan sample and concurred that the type is G H P 76 anchor. It is approximately the prevalent in the region (.I. Guba, 1998, same size, which would preclude its use pers. comm.). The broken G H P 3 1 and for anything more than as a lightweight four-hole G H P 32 anchors are sedimenanchor for a small fishing vessel or lighter. tary. G H P 31 was probably formed from The Kommos example has been recon- beach sand that is common to many structed as a.triangular one-holed anchor, coastal areas and makes the provenance 338
C. SOUTER: STONE ANCHORS
difficult to determine. GHP 32 is also a beach rock made up of shell material and quartz. This type of stone is common to the Arabian coast from Abu Dhabi to Kuwait (Green, 1998: 35).
Discussion GHP 30 and GHP 31 anchors are categorised as Arab-Indian, the name reflecting the extensive geographic provenance of such anchor types. Similar examples are known from sites in the Western Indian Ocean--Mombasa, Malindi, Minicoy, Oman, Yemen, Iran, India and the Red Sea. There are about 80 anchors of the Arab-Indian type catalogued in the Indian Ocean and the Red Sea. The largest underwater concentrations have been found off the coast near Qalhat, Oman and at Dwarka, India (Vosmer, 1998, pers. comm.). The Omani find consists of approximately 27 anchors or anchor fragments (Vosmer, 1998, pers. comm.). Twenty-four similar examples have been reported in India. However, all were being reused as architectural elements in dock works and forts (Tripati & Gaur, 1997). There has been a wide range of dates presented for this type of anchor. Some scholars categorise them as Late Bronze Age (Frost, 1963; Nibbi, 1991). Generally they are termed Arab-Indian and attributed to the period 6th to 16th century AD. Attempts have been made to compare them to earlier anchors such as the Temple of Isidorus finds from the Alexandrian Delta which were considered to come from Roman levels by Frost (Owen, 1997: 5). However, in Frost's Criteria for a Corpus (Frost, 1986) the anchor's form is neglected in such a comparison. Whereas the Roman example shares some common diagnostic features, it is markedly different in shape and certainly not an Arab-Indian type. The Galle Harbour anchors come from an unstratified site, which prevents absolute dating. The Arab-Indian anchor
is best understood by a comparative stylistic analysis of similar finds. The Arab-Indian anchor as interpreted by Raban was shaped for a safer anchorage in a coral seabed, like those fringing the Indian Ocean. It is believed that large vessels would carry a complement of anchors that would be used simultaneously to secure the ship (Frost, 1983:357) or replace those lost. It is conceivable that several anchor types were carried to be used in conjunction with each other. The sheer size of these Arab-Indian anchors suggests that they would not be easily manoeuvred and deployed. This may have contributed to their loss at sea and perhaps to the cluster in Galle Harbour. By contrast, this site is known as a traditional mooring ground. This supports the idea that large Arab-Indian anchors here were not intended for use as working ships anchors, that is to be cast from a ship, but rather as mooring anchors. It may explain the deployment of G H P 30, with an estimated weight of 1000kg. Its position inshore gives credence to the idea that mooring anchors were laid carefully from a floating platform either using a pulley system (Kapit/in, 1994: 4) or a boom method as depicted on an often cited 8th-century Cypriot vase (Frost, 1970: 392). A series of fixed moorings in the Kamba Bandina Gala area would have provided ships with a useful anchorage within the harbour. The system of fixed moorings may have been imported by traders, as there were generally very few harbour facilities in the western Indian Ocean (Kapit~in, 1994: 4). However, this type of anchor, regardless of size, was still probably intended and used as an open water ship's anchor. The weight of such an anchor would enable it to be used for retarding a vessel mid-water effectively (Nibbi, 1993: 11). The difficulties associated with the physical management of such a large object would not have precluded its use in the open sea. 339
NAUTICAL ARCHAEOLOGY, 27.4
The suggestions that G H P 30 was either a fixed mooring or a working ship's anchor are both plausible. G H P 31 and the broken anchor left in situ may have been used as mooring anchors after they were broken, and in this context are example.s of secondary use. However, it is also possible that they were simply abandoned when local replacements were obtained. The Arab-Indian type anchors would be more likely to break when being deployed or recovered. A comprehensive analysis of the relationship of the anchors to one another as they lie on the seabed, over a number of similar sites, may begin to reveal whether the in situ configuration denotes a formal mooring anchorage. The suggestion of G H P 32 as a composite anchor has recently been evidenced with the discovery of wood remains within a three-hole stone anchor from Crusader Arsuf (Appollonia) in Israel (Grossman & Kingsley, 1996: 51). W o o d remains were discovered in one of the lower holes. The wood had been sawn carefully to fit the holes and sat flush with the surface of the anchor on the upper and presumably under side. The condition and position of the wood in situ suggests that the anchor was deployed without the envisioned large wooden flukes protruding well beyond the surface of the stone (Grossman & Kingsley, 1996: 52). This find would therefore support the idea that composite anchors in some circumstances were used without the arms complete. G H P 32 also shares characteristics with examples found in India in relation to the hole configuration and general shape. The Indian examples, in particular those -from Sindhudurg Fort on the West Coast are, however, more triangular in form. The date range for this form in India is from 2300 BC to the Historic Period. (Tripati & Gaur, 1997: 55). Conclusion The anchor assemblage found in Sri Lanka supplements a burgeoning corpus of stone 340
anchors found in the Indian Ocean region. The area containing the anchor site is a formal mooring 9 r lightering site for Galle. The discovery of these various anchor types indicates a continuity of use of this area. The variety of anchors discovered in close proximity to each other in Galle Harbour leads to the conclusion that a specific type of anchor was selected to suit mariners' requirements, the ability of a ship to transport the particular anchor type and the availability of materials. It is also conceivable that all these types of anchors were used contemporaneously. The position of anchors in situ can indicate sea-lanes, the navigational style of seafarers (Frost, 1977: 378) and possible ancient anchorages. In classifying an anchor, there is a tendency to compare typological features, which attribute the style to a particular region or culture. The Galle H a r b o u r anchors suggest that the identification of an anchor type or style in a geographical sense does not necessarily indicate its provenance. The possible function of larger examples, such as G H P 30 as moorings, also hinders the attribution, on purely typological features, of an anchor style to a particular vessel. Anchors to some extent were consumables and would need to be replaced in foreign ports. Anchors were also trade items. K a m b a Bandina Gala is a known mooring area and this accumulation of anchors contributes to our understanding of it as an anchorage site. The recommendations for further archaeological work on the anchor site are dependent on the development and installation of a conservation management plan, thereby ensuring that these artefacts are properly preserved. The timbers, G H P 44 and G H P 4 5 , are being treated at the Western Australian Maritime Museum as part of the Centre of Excellence maritime archaeological conservation programme and will then be returned to Sri Lanka for display along with the stone anchor shank
C. SOUTER: STONE ANCHORS
GHP 30, with which they were found. A further examination of the anchor site, as well as the surrounding area of Kamba Bandina Gala, is also required to establish if there are any more related artefacts. The area is littered with material, especially along the reef which runs through the site. A more thorough survey of this reef and the material around it may shed light on the context in which the stone anchors were found. It is recommended that a comparison of similar anchorages in the region be undertaken. A formal sampling of similar stone anchor sites in the Indian Ocean region would enable us to determine and distinguish various typologies, their frequencies and the nature of the site and condition of
the anchors. A further analysis related to their function based on the physical degree of wear may also be undertaken.
Acknowledgements The Galle Harbour Project would like to acknowledge in Sri Lanka: the Director General of Archaeology, Dr S. U. Deraniyagala; Director Post Graduate Institute of Archaeological Research, Prof. Senake Bandranayake; Lt Cdr Somasiri Devendra; Mr David Ritchie, Australian High Commission Sri Lanka; Sri Lankan Sub Aqua Club; Sri Lanka Navy; Mr D. Kandamby, National Maritime Museum Galle; Mr H. A. Gunatilaka, Sri Lankan Geology Consultant.
References Bolza & Keating, 1982, Characteristics, Properties and Uses of Timbers. South East Asia, North Australia and the Pacific. Incarta Press. Carpenter, J., 1993, Conservation h7 Galle. Department of Materials Conservation. Western Australian Maritime Museum. Chittick, N., 1980, Stone anchor shanks in the Western Indian Ocean. IJNA, 9: 73-76. Frost, H., 1963, Under the Mediterranean. Marine Antiquities. London. Frost, H., 1970, Bronze Age anchors from the Eastern Mediterranean. MM, 56: 377-376. Frost, H., 1979, Egypt and stone anchors: some recent discoveries. MM, 65: 137-161. Frost, H., 1982, The birth of the stocked anchor and the maximum size of ships. MM, 68: 263-278. Frost, H., 1986, Stone anchors: criteria for a corpus. Thracia Pontica III: 352-369. Frost, H., 1993, Stone anchors: a reassessment reassessed. MM, 79: 449-458. Galili, E., 1993, Underwater surveys and rescue excavations along the Israeli coast. IJNA, 22: 61-77. Green, J. 1971, An underwater archaeological survey of Cape Andreas, Cyprus, 1969-70: a preliminary report. Proceedings of the 23rd Symposium of the Colston Research Society. Vol XXIII. London. Green. J. (Ed.), 1998, Sri Lanka Department of Archaeology Report on the Joint Sri Lanka-AustraliaNetherlands Galle Harbour Project 1996-1997. Special Publication No. 4, Australian National Centre of Excellence for Maritime Archaeology. Grossman, E. & Kingsley, S., 1996, A three-hole stone anchor with wood remains from Crusader Arsuf (Apollonia), Israel. IJNA, 25: 49-54. Kapit~in, G., 1994, Stone-shank anchors of the Arab-Indian trade period--were they mooring anchors? AlMA Bulletin, 18: 1-6. Moll, A., 1927, History of the anchor. MM, 13: 4. Nibbi, A., 1991, Five stone anchors from Alexandria. IJNA, 20: 185-194. Nibbi, A., 1992, A group of stone anchors from Mirgissa on the upper Nile. IJNA, 21: 259-267. Nibbi, A., 1993, Stone anchors: the evidence re-assessed. MM. 79: 5-26. Owen, J., 1997, Do Anchors Mean Ships? Underwater evidence for maritime trade along the Dofar coast, Southern Indian Ocean. In: Avanzini, A. (Ed.) Profilmi d'Arabia. 351-363. Raban, A., 1990, Medieval anchors from the Red Sea. 1JNA, 19: 299-306. Shaw, J., 1995, Two three-holed stone anchors from Kommos, Crete: their context, type and origin. IJNA, 24: 279-291. Tripati, S. & Gaur, A., 1997, Stone anchors from Sindhudurg Fort on the West Coast of India. IJNA, 26: 51-57. 341
NAUTICAL ARCHAEOLOGY, 27.4 Vosmer, T., 1997, Oman Expedition Field Report. Western Australian Maritime Museum Report No. 130. Vosmer, T., 1998, Stone Anchors of the Western Indian Ocean and Red Sea http:llwww.mm.wa.gov.aul Museum/march/stoneanchors/Stone_Anchors.html
342
Stone anchors near Black Fort, Galle,. Sri Lanka
®
Corioli Souter Australian National Centre of Excellence for Maritime Archaeology, Western Australian Marithne Museum, Cliff Street, Fremantle WA 6160, Australia
Introduction The Joint Australia-Sri LankaNetherlands Maritime Archaeological Project is centred in the World-Heritagelisted Galle Harbour, Sri Lanka. The harbour is notable for the VOC (Dutch East India Company) shipwrecks it contains and has been the subject of underwater archaeological surveys since 1992. Galle was an operational port and major entrepot in pre-Christian times with the area coming to prominence in the 12th century with the arrival of the Portuguese and later the Dutch. Galle was the second most important harbour of the VOC, strategically located in terms of trade between the Arabian Peninsula and East Asia. Galle remained the major port in Sri Lanka after the arrival of the British in 1796 and it was only in 1873 when the present port of Colombo was constructed that Galle became marginalised. During the 1996 season, a side scan sonar survey (Green, 1998) led to the discovery of a stone anchor site. A S C U B A circular search to investigate a side scan anomaly resulted in the discovery of one large Arab-Indian stone anchor, G H P 30 (Fig. 1), and a broken portion from a smaller anchor of the same type. The Arab-Indian anchor type has been dated to the period from the 6 t h t o 16th century A D (Vosmer, 1998, pers. comm.). ,The larger anchor was recorded as 350 cm long, the broken anchor approximately 150 cm. Both were located close to the Black Fort, 1057-2414/98/040331 + 12 $30.0010
in the area of K a m b a Bandina Gala, a traditional mooring site for vessels lightering cargo to and from Galle. In the same general area is a site known to have ceramics from a wide geographical and temporal range. This could possibly be associated with the mooring of vessels and the stone anchors. When discovered, the anchors lay flat and half-buried on a sand seabed covered with a layer of fine silt and organic material. During the 1997 season it was planned to raise these stone anchors and to determine whether there were any inscriptions; for petrological studies to ascertain the origin of the stone; and conservation. Following their conservation it was intended that the anchors be incorporated into a display for the Maritime Museum in the Galle Fort.
Site description The Galle Harbour seabed consists of a mixture of sand, sediment and rock overlying mud deposits. A small limestone reef runs through the middle of the anchor site. At the time of survey the underwater visibility was extremely poor, ranging between 0 m to 1"5 m at best. The visibility improved in the week prior to full moon, indicating that it was influenced by tidal and subsequent sediment movement. Sri Lankan Sub Aqua Club divers, having made underwater observations in the harbour following the monsoon season, reported considerable upheaval of the seabed resulting in exposure of shipwrecks © 1998 The Nautical Archaeology
Society
N A U T I C A L
A R C H A E O L O G Y ,
~----;< ~;
"
. i..:'.:..,~ ':'i
27.4
• ."::~'
.'l
...
I
•
~
"
'~
L
.. ?'.
.
..
"- -": "-;L: '
I
-,-..'-!-.:~.....-?:~-:,T:.;.:...::.r..,.,,,,.~.x,..~:~.....,~,~.~....~.. • • , , . . . '
; !!!:;v:: : • . .
"
" .';*~;
.
.
" "•
.
'
". '
..r,
.
.... ". ""."." "
"
•
..
. . . .
' ' : " i!i':
I
I
0
50
" • "-~:'~,."', r " ; ~ ~ ' "'
•
,~....:..~.-~ ,.. :
": . . . .
•
. . . .
...~
"
•
..~
.
' ~
- . - . : , : = . . . ~
......:i::i:"i
.....
c m
Fig~tre 1. Line drawing of GHP 30.
and dispersal of artefacts. The anchor site is approximately 7 m deep and is interspersed with broken ceramics, stone fragments, iron concretions, firebox slag, bones, coal and modern rubbish. Although the area is subject to surge, the seabed is generally level except where the presence of large objects, such as the stone anchors, has resulted in scour or build-up of sediment on one face. Search
and
survey
During the 1997 season, diver search methods were employed to relocate the anchors. Circular jackstay searches and swim lines were used to ensure systematic coverage of the seabed in low visibility, to locate the objects. The large stone anchor GHP 30 was relocated first. Another stone object, approximately 50cm in diameter, was found 14 m from G H P 30. The round hole 332
in this stone suggests that it might be another broken portion of an anchor. The object was identified as a modern mooring upon recovery, with a small, machine drilled, round hole and a modern piece of chain concreted to the surface of the rock. During this search, a third previously unrecorded stone anchor, G H P 32, was found. It was a flat trapezoidal anchor with two round vertically- aligned holes at the top and two square horizontallyaligned holes located in the lower half (Fig. 2). While the two round holes present an unusual configuration, the shape is typical of the Mediterranean. The search continued m an easterly direction from this anchor in an attempt to locate the broken anchor found in 1996 (Fig. 3). A crown of an Arab-Indian anchor which had broken off at the second square hole from the top was discovered.
C. SOUTER: STONE ANCHORS
Exposed side
Underside
I
50
0 cln
Figure 2. Line drawing of G H P 32.
I ?
, I
' I I
I I
', [
) Profile
I
I
0
I
50
cm Figure 4. Line drawing of G H P 31. 0
5O
Figure 3. Line drawing of 1996 broken anchor in situ.
However, the dimensions of this anchor, 74 cm x 37 cm, were smaller than those of the broken anchor recorded during the 1996 expedition This suggests that this was a new find and not the original example being sought. The difficulty in identification was naturally compounded by the poor visibility. This new anchor was recorded and labelled G H P 31 (Fig. 4).
During searches south of the large anchor G H P 3 0 ' s position, a round anchor, GHP 76, with a diameter of 65 cm and height of 15 cm was also discovered. G H P 76 is essentially a rounded stone with a squared off hole through the centre (Fig. 5). An 18th-century admiralty iron anchor was also found in this area. The relative positions for each anchor, the modern mooring and the rock outcropping from the seabed were recorded by trilateration. From these anchor positions subsequent circular searches could be used 333
NAUTICAL ARCHAEOLOGY, 27.4
I 0
cm
I 50
Figure 5. Line drawing of G H P 76.
to locate further artefacts and to better define the site by location and mapping of the natural features. All positions were incorporated into the Rule Web program and the resultant co-ordinates plotted into a site plan (Fig. 6) Anchor recovery
In preparation for raising, rope was tied around the largest anchor, G H P 30, at three intervals along the length of the shank. A space under the anchor G H P 30 was cleared using a water jet in order to attach the ropes. During this procedure a substantial piece of blackened timber was located. It was thought that this wood could be a remnant of the anchor arm or stock. Further excavation of the area under G H P 30 revealed two pieces of hardwood. Hardwood timber would be strong and durable enough for anchoring purposes. The larger piece GHP 45 has been identified as belonging to the Leguminosae family, best matching the Xylia species (pyinkado), common in tropical Asia. The species is known as irul in India which suggests its presence on the sub-continent and possibly Sri Lanka. The smaller timber, GHP44, was also identified as belonging to the Leguminosae family and is most likely to be Sindora species (sepitir, sindoer). This species is native to South East Asia. There is no reference to the species being indigenous to Sri Lanka or 334
India (Bolza & Keating, 1982). This is the first time wood has been found in association with an Arab-Indian stone anchor. It is hoped that the discovery of timber will provide insight into the configuration of these anchors. The timber's position in relation to the stone shank and in particular the square holes suggests that the wood functions as the arms of the anchor. The first piece raised, G H P 4 4 (Fig. 7), is 980 mm in length, 160 mm at its widest point and 104ram high. It was found approximately 300mm under the sediment. There is no uniform stratigraphy as the first layers of sediment contain a random mix of broken ceramics and roofing tiles. The lack of stratigraphy implies that the area is, or has been, subject to a high degree of disturbance. The uppermost surface of the timber as found & situ exhibits curvature with most of the original surface intact. The surface also exhibits tool marks, probably those associated with adzing. The underside has been partially degraded by marine borers although much of the original surface remains. The original shape of the timber has been retained and has a natural concave curve on one edge approximately 73 cm along its length. It appears as though the curve of the timber has been tooled and the edges bevelled to shape an arm/stock. It is difficult to ascertain the original shape at this end of the timber as a large section is missing. However, in profile view the end appears originally to have been finished in a chisel point. With this in mind, the missing portion of timber may have been symmetrical with that which remains, together forming a long, gradually tapered, V-shaped anchor arm or possibly stock. The thicker end of the timber, which was found next to the shank, is mortised and bears extensive degradation by marine borers. It appears that the mortised end has snapped off suggesting that this piece is only a portion of the original timber. The function of the mortised end is unclear. It
C. SOUTER: STONE ANCHORS
Modern
Mooring @
'IN
GI-IP 30
GHP32
* GHP31 Broken anchor in situ
:~
~Iron
anchor in situ
GHP 76
I
50
lIl
Figure 6. Site plan.
0
50 cm
Figure 7. Line drawing of GHP 44.
may have performed as a locking mechanism to secure the timber into the anchor shank. The second piece raised, G H P 45, is considerably larger than the first (Fig 8). It is 134 cm in length, 26 cm wide and up to 18 cm high. The timber is similar to the smaller piece as it exhibits a natural curve, which has been enhanced by tooling. The
I
1
0
50
cm
Figure 8. Line drawing of GHP 45.
timber has retained its original surface, in part. Raised knots, up to 5 m m high on the upper surface of the wood as found in situ, may indicate where the original surface was and the subsequent extent of degradation, although it is also conceivable that they are typical wood growth formations and part of the timber surface. The grain of the timber is visible on the surface of the 335
NAUTICAL ARCHAEOLOGY, 27.4
thicker end. This end lay perpendicular to the anchor, under the stone shank just behind the crown. It is possible that either it has broken off from another piece of timber or it has broken off from the hole in the anchor shank, as it was excavated from beneath the anchor. The curve of the timber begins 65 cm from the wider end to form what may be interpreted as an arm. The timber may have been selected for its naturally contoured shape and then mechanically worked to create an arm, which would penetrate and grip into the sediment. The end appears to finish in a point, although this, too, is broken or worn. Tool marks are visible particularly in profile, although not to the same extent as the smaller timber. The underside of the large timber, as found in situ, shows evidence of degradation due to marine borers with some of the original surface remaining. It curves upward contributing to the hooked-arm appearance of the timber. Timber samples were collected for identification, CI4 dating and dendrochronological analysis. The radiocarbon age determination for the sample from GHP44 was 4 3 0 + 8 0 B P , by The University of Waikato Radiocarbon Dating Laboratory (1998). The timbers as we find them are incomplete, which prevents overall size determination. Both pieces are too small to fit well within the square holes (Fig. 9). It is also likely that the timber has swelled since immersion. If this were the case the arms may have been secured in the holes by a wedge or tenons as suggested by the mortised end of G H P 44.
Stone anchor descriptions G H P 3 0 is made from a sedimentary polymict conglomerate and is 317 cm in length. It is 55 cm at the crown and tapers to 26 cm at the opposite end. G H P 31 is composed of a sedimentary sandstone beachrock and has much smaller dimensions being 74 cm long. It is 30 cm wide at 336
Figure 9. GHP 30 with timbers. (Photo: Patrick Baker)
the hawser end and a maximum of 37 cm at the break, approximately half-way up the shank. The anchors are oblong in shape with a round hole in one end, presumably to take the mooring rope, and two square holes in the other end set at 90 ° to each other, one slightly further up the shank than the other. Through these opposed facing holes wooden stakes were set to act as the arms or the stock of the anchor. The type was initially identified by Honor Frost as a grapnel due to the transverse square holes in the lower part of the shank. This has been disputed as these square holes are different sizes and the anchor shanks are usually rectangular in cross section while the shafts of fourarmed grapnels are square or round in cross section (Kapit/in, 1994: 2). They are also somewhat smaller in length. Kapitfin has proposed a reconstruction where the larger hole near the end of the shank actually holds a stock rather than an arm (Kapit/in, 1994: 2). The discovery of wood in association with G H P 30 suggests a different reconstruction. The in situ positions of the two baulks of wood in relation to
C. SOUTER: STONE ANCHORS
the anchor holes indicate that the arm came from the hole nearest the crown. The position of the timbers on the seabed suggests that the lower hole on the stone shank held a curved wooden arm and the upper hole held a smaller piece of wood, which may be either an arm or stock. The round hole at the other end of the shank would have had a rope directly tied through it or perhaps another stock. The suggestion of a stock through this hole is evidenced by other anchor examples such as a shank found in the Red Sea at Lone Mushroom, west of Ras Muhammed (Raban, 1990: 302). This Lone Mushroom anchor is rectangular in cross section but only has one rectangular hole in the lower part that passes at right angles to the axis of the circular hole at the other end. This arrangement suggests that the rectangular hole held the arm of the anchor while the round hole held the stock. A timber would be required in the round hole in order for the anchor to grip into the seabed effectively (Kapit~in, 1994: 2). Alternatively, the notches at the base of the crown on the Lone Mushroom anchor may have been used to lash another timber to the base to be used as an arm/stock, or they may simply be quarry marks. However, the size of the circular holes in the Sri Lankan examples, suggests that only a rope would be passed through them. G H P 3 2 is a four-hole composite anchor, which essentially means that the anchor does not rely on its weight alone (Nibbi, 1991: 191). It is constructed from sedimentary sandstone and is 100 cm high. The maximum width is 83 cm tapering into 40 cm near the top. The anchor has a thickness of 17 cm. The type has been labelled Byzantine-Arab, as they were common in this period (Frost, 1963: 49). They were still in use until recently, both in the Persian Gulf and more commonly the Mediterranean, and as a result have also been called Mediterranean in form (Frost, 1963: 13).
G H P 32 has a pock-marked surface as a resuli~ of exposure and encrustation. Only a small portion of the original surface remains. The anchor has two round holes below the apex and two square holes along the base. It is generally accepted that the top hole(s) in the anchor, were for the main hawser. This is an unusual example as this style of anchor as found in other contexts has only one round hole below the apex. This may be a deliberate double-hole configuration to ensure the anchor stone is secured to the lifting rope, or perhaps the lower hole was added later to replace the original hole which is only 45 mm from the top of the anchor. If the original hole was used on its own it may have caused the stone to break under strain. The anchor has random channels cut into the surface, possibly wear marks from rubbing of the hawser rope or ropes on the stone surface while the anchor was tethered on the sea floor or to the ship during transportation. The channels may have resulted from lashing the timber arms to the anchor and subsequent wear by the rope. Alternatively, the channels may be the result of environmental conditions affecting the stone in situ such as sand built up on the face of the anchor scouring out the sandstone. The square and round holes also show signs of wear. G H P 76 is a round, stone anchor with a single, squared perforation. It is constructed from sedimentary sandstone with the dimensions 65 x 49 x 15 cm. This round anchor is possibly a weight anchor• This type relies on its weight as opposed to arms to hold a vessel. Its shape and size may indicate that it was used for a small vessel on a rocky sea floor (Vosmer, 1997: 9). The purpose of the hole's square finish is unclear• The shape of the hole is squared off perhaps so that a wooden arm or stock may be inserted and locked into place. In comparison, the holes in the Late Bronze Age Ugaritic anchors (Frost, 1970: 388) are cupular drilled yet chiselled square on 337
NAUTICAL ARCHAEOLOGY, 27.4
one side perhaps to hold the wood in place which was common to the Minoan Period and stop it loosening through rotation. (Nibbi, 1993: 14). I f G H P 76 is a fragment However, in the Ugaritic anchors this tech- of a larger anchor it is also conceivable nique was thought to be decorative as the that it may be the remains of an A r a b original drill holes are narrow and round Indian type. The size of the anchor in and not cut to fit square flukes (Frost, cross-section appears to negate this, being 1970: 388). There is no evidence to support marginally smaller than other known exthis in other single hole anchors. However, amples. It is not clear, however, how fragsquare holes in the Mediterranean-style mentary this example is. If the anchor is anchors have generally denoted the use of classed as Arab-Indian it would also better wooden arms. It is not necessary to square explain the square hole. In terms of function of an anchor, it is off a hole if it is to be used for a hawser alone. The'shape of the anchor has some the chain or rode as opposed to the anchor similarities to the ringstones used in the weight itself which holds a ship in position Maldives (Vosmer, 1997: 17). Ringstone (Green, 1971: 175). G H P 7 6 may have anchors were used for anchoring on coral been used in conjunction with several simireefs to obviate snagging on the coral. A lar one-hole anchor types to weigh down forked branch was inserted through the the anchor rode to allow it to hang parallel hole of a ringstone operating as a crude with the sea-bed. With this system, a cbmshank. (Vosmer, 1997: 17). The ringstone posite anchor would be attached to the end anchor comparison, however, fails to of the rode with a number of one-hole address where G H P 76 was found. The anchors attached to the rope at intervals sand and sediment seabed of Galle (Green, 1971: 175). This type of simple Harbour is unsuitable for either rock or single hole anchor is also in use today and coral anchor. Its function as found in situ is makes dating G H P 76 difficult. probably explained as a temporary mooring or abandoned anchor for a smaller Petrological analysis boat. It is also possible that it is not in its Most stone local to Sri Lanka is igneous. original form. It may be a fragment, such The stone sample from the large anchor G H P 30 has been identified as a sedimenas the lower corner of a Mediterranean three-hole anchor, similar to G H P 32. This tary conglomerate by Sri Lankan geolowould explain the square hole and the gist, A. Guntalika (Green, 1998: 35). This unusual shape. The closest comparison to type of stone is quite common in alluvial this anchor comes from Kommos in fan or fan delta complexes flanking rising Southern Crete and comes from a defin- mountain fronts or fault scarps. From able chronological context (Shaw, 1995: initial inspection the the most probable 282). It was dated to the late Minoan source of the anchor stone is Oman, as Period of 1250 BC (Nibbi, 1993: 13). The similar conglomerate formations occur Kommos example, artefact Number $636, along the mountain front running along is one of five anchor fragment finds and it the Omani coastline (Green, 1998: 35). A exhibits the same characteristics including geologist in Oman also examined the the squared off hole, as the Sri Lankan sample and concurred that the type is G H P 76 anchor. It is approximately the prevalent in the region (.I. Guba, 1998, same size, which would preclude its use pers. comm.). The broken G H P 3 1 and for anything more than as a lightweight four-hole G H P 32 anchors are sedimenanchor for a small fishing vessel or lighter. tary. G H P 31 was probably formed from The Kommos example has been recon- beach sand that is common to many structed as a.triangular one-holed anchor, coastal areas and makes the provenance 338
C. SOUTER: STONE ANCHORS
difficult to determine. GHP 32 is also a beach rock made up of shell material and quartz. This type of stone is common to the Arabian coast from Abu Dhabi to Kuwait (Green, 1998: 35).
Discussion GHP 30 and GHP 31 anchors are categorised as Arab-Indian, the name reflecting the extensive geographic provenance of such anchor types. Similar examples are known from sites in the Western Indian Ocean--Mombasa, Malindi, Minicoy, Oman, Yemen, Iran, India and the Red Sea. There are about 80 anchors of the Arab-Indian type catalogued in the Indian Ocean and the Red Sea. The largest underwater concentrations have been found off the coast near Qalhat, Oman and at Dwarka, India (Vosmer, 1998, pers. comm.). The Omani find consists of approximately 27 anchors or anchor fragments (Vosmer, 1998, pers. comm.). Twenty-four similar examples have been reported in India. However, all were being reused as architectural elements in dock works and forts (Tripati & Gaur, 1997). There has been a wide range of dates presented for this type of anchor. Some scholars categorise them as Late Bronze Age (Frost, 1963; Nibbi, 1991). Generally they are termed Arab-Indian and attributed to the period 6th to 16th century AD. Attempts have been made to compare them to earlier anchors such as the Temple of Isidorus finds from the Alexandrian Delta which were considered to come from Roman levels by Frost (Owen, 1997: 5). However, in Frost's Criteria for a Corpus (Frost, 1986) the anchor's form is neglected in such a comparison. Whereas the Roman example shares some common diagnostic features, it is markedly different in shape and certainly not an Arab-Indian type. The Galle Harbour anchors come from an unstratified site, which prevents absolute dating. The Arab-Indian anchor
is best understood by a comparative stylistic analysis of similar finds. The Arab-Indian anchor as interpreted by Raban was shaped for a safer anchorage in a coral seabed, like those fringing the Indian Ocean. It is believed that large vessels would carry a complement of anchors that would be used simultaneously to secure the ship (Frost, 1983:357) or replace those lost. It is conceivable that several anchor types were carried to be used in conjunction with each other. The sheer size of these Arab-Indian anchors suggests that they would not be easily manoeuvred and deployed. This may have contributed to their loss at sea and perhaps to the cluster in Galle Harbour. By contrast, this site is known as a traditional mooring ground. This supports the idea that large Arab-Indian anchors here were not intended for use as working ships anchors, that is to be cast from a ship, but rather as mooring anchors. It may explain the deployment of G H P 30, with an estimated weight of 1000kg. Its position inshore gives credence to the idea that mooring anchors were laid carefully from a floating platform either using a pulley system (Kapit/in, 1994: 4) or a boom method as depicted on an often cited 8th-century Cypriot vase (Frost, 1970: 392). A series of fixed moorings in the Kamba Bandina Gala area would have provided ships with a useful anchorage within the harbour. The system of fixed moorings may have been imported by traders, as there were generally very few harbour facilities in the western Indian Ocean (Kapit~in, 1994: 4). However, this type of anchor, regardless of size, was still probably intended and used as an open water ship's anchor. The weight of such an anchor would enable it to be used for retarding a vessel mid-water effectively (Nibbi, 1993: 11). The difficulties associated with the physical management of such a large object would not have precluded its use in the open sea. 339
NAUTICAL ARCHAEOLOGY, 27.4
The suggestions that G H P 30 was either a fixed mooring or a working ship's anchor are both plausible. G H P 31 and the broken anchor left in situ may have been used as mooring anchors after they were broken, and in this context are example.s of secondary use. However, it is also possible that they were simply abandoned when local replacements were obtained. The Arab-Indian type anchors would be more likely to break when being deployed or recovered. A comprehensive analysis of the relationship of the anchors to one another as they lie on the seabed, over a number of similar sites, may begin to reveal whether the in situ configuration denotes a formal mooring anchorage. The suggestion of G H P 32 as a composite anchor has recently been evidenced with the discovery of wood remains within a three-hole stone anchor from Crusader Arsuf (Appollonia) in Israel (Grossman & Kingsley, 1996: 51). W o o d remains were discovered in one of the lower holes. The wood had been sawn carefully to fit the holes and sat flush with the surface of the anchor on the upper and presumably under side. The condition and position of the wood in situ suggests that the anchor was deployed without the envisioned large wooden flukes protruding well beyond the surface of the stone (Grossman & Kingsley, 1996: 52). This find would therefore support the idea that composite anchors in some circumstances were used without the arms complete. G H P 32 also shares characteristics with examples found in India in relation to the hole configuration and general shape. The Indian examples, in particular those -from Sindhudurg Fort on the West Coast are, however, more triangular in form. The date range for this form in India is from 2300 BC to the Historic Period. (Tripati & Gaur, 1997: 55). Conclusion The anchor assemblage found in Sri Lanka supplements a burgeoning corpus of stone 340
anchors found in the Indian Ocean region. The area containing the anchor site is a formal mooring 9 r lightering site for Galle. The discovery of these various anchor types indicates a continuity of use of this area. The variety of anchors discovered in close proximity to each other in Galle Harbour leads to the conclusion that a specific type of anchor was selected to suit mariners' requirements, the ability of a ship to transport the particular anchor type and the availability of materials. It is also conceivable that all these types of anchors were used contemporaneously. The position of anchors in situ can indicate sea-lanes, the navigational style of seafarers (Frost, 1977: 378) and possible ancient anchorages. In classifying an anchor, there is a tendency to compare typological features, which attribute the style to a particular region or culture. The Galle H a r b o u r anchors suggest that the identification of an anchor type or style in a geographical sense does not necessarily indicate its provenance. The possible function of larger examples, such as G H P 30 as moorings, also hinders the attribution, on purely typological features, of an anchor style to a particular vessel. Anchors to some extent were consumables and would need to be replaced in foreign ports. Anchors were also trade items. K a m b a Bandina Gala is a known mooring area and this accumulation of anchors contributes to our understanding of it as an anchorage site. The recommendations for further archaeological work on the anchor site are dependent on the development and installation of a conservation management plan, thereby ensuring that these artefacts are properly preserved. The timbers, G H P 44 and G H P 4 5 , are being treated at the Western Australian Maritime Museum as part of the Centre of Excellence maritime archaeological conservation programme and will then be returned to Sri Lanka for display along with the stone anchor shank
C. SOUTER: STONE ANCHORS
GHP 30, with which they were found. A further examination of the anchor site, as well as the surrounding area of Kamba Bandina Gala, is also required to establish if there are any more related artefacts. The area is littered with material, especially along the reef which runs through the site. A more thorough survey of this reef and the material around it may shed light on the context in which the stone anchors were found. It is recommended that a comparison of similar anchorages in the region be undertaken. A formal sampling of similar stone anchor sites in the Indian Ocean region would enable us to determine and distinguish various typologies, their frequencies and the nature of the site and condition of
the anchors. A further analysis related to their function based on the physical degree of wear may also be undertaken.
Acknowledgements The Galle Harbour Project would like to acknowledge in Sri Lanka: the Director General of Archaeology, Dr S. U. Deraniyagala; Director Post Graduate Institute of Archaeological Research, Prof. Senake Bandranayake; Lt Cdr Somasiri Devendra; Mr David Ritchie, Australian High Commission Sri Lanka; Sri Lankan Sub Aqua Club; Sri Lanka Navy; Mr D. Kandamby, National Maritime Museum Galle; Mr H. A. Gunatilaka, Sri Lankan Geology Consultant.
References Bolza & Keating, 1982, Characteristics, Properties and Uses of Timbers. South East Asia, North Australia and the Pacific. Incarta Press. Carpenter, J., 1993, Conservation h7 Galle. Department of Materials Conservation. Western Australian Maritime Museum. Chittick, N., 1980, Stone anchor shanks in the Western Indian Ocean. IJNA, 9: 73-76. Frost, H., 1963, Under the Mediterranean. Marine Antiquities. London. Frost, H., 1970, Bronze Age anchors from the Eastern Mediterranean. MM, 56: 377-376. Frost, H., 1979, Egypt and stone anchors: some recent discoveries. MM, 65: 137-161. Frost, H., 1982, The birth of the stocked anchor and the maximum size of ships. MM, 68: 263-278. Frost, H., 1986, Stone anchors: criteria for a corpus. Thracia Pontica III: 352-369. Frost, H., 1993, Stone anchors: a reassessment reassessed. MM, 79: 449-458. Galili, E., 1993, Underwater surveys and rescue excavations along the Israeli coast. IJNA, 22: 61-77. Green, J. 1971, An underwater archaeological survey of Cape Andreas, Cyprus, 1969-70: a preliminary report. Proceedings of the 23rd Symposium of the Colston Research Society. Vol XXIII. London. Green. J. (Ed.), 1998, Sri Lanka Department of Archaeology Report on the Joint Sri Lanka-AustraliaNetherlands Galle Harbour Project 1996-1997. Special Publication No. 4, Australian National Centre of Excellence for Maritime Archaeology. Grossman, E. & Kingsley, S., 1996, A three-hole stone anchor with wood remains from Crusader Arsuf (Apollonia), Israel. IJNA, 25: 49-54. Kapit~in, G., 1994, Stone-shank anchors of the Arab-Indian trade period--were they mooring anchors? AlMA Bulletin, 18: 1-6. Moll, A., 1927, History of the anchor. MM, 13: 4. Nibbi, A., 1991, Five stone anchors from Alexandria. IJNA, 20: 185-194. Nibbi, A., 1992, A group of stone anchors from Mirgissa on the upper Nile. IJNA, 21: 259-267. Nibbi, A., 1993, Stone anchors: the evidence re-assessed. MM. 79: 5-26. Owen, J., 1997, Do Anchors Mean Ships? Underwater evidence for maritime trade along the Dofar coast, Southern Indian Ocean. In: Avanzini, A. (Ed.) Profilmi d'Arabia. 351-363. Raban, A., 1990, Medieval anchors from the Red Sea. 1JNA, 19: 299-306. Shaw, J., 1995, Two three-holed stone anchors from Kommos, Crete: their context, type and origin. IJNA, 24: 279-291. Tripati, S. & Gaur, A., 1997, Stone anchors from Sindhudurg Fort on the West Coast of India. IJNA, 26: 51-57. 341
NAUTICAL ARCHAEOLOGY, 27.4 Vosmer, T., 1997, Oman Expedition Field Report. Western Australian Maritime Museum Report No. 130. Vosmer, T., 1998, Stone Anchors of the Western Indian Ocean and Red Sea http:llwww.mm.wa.gov.aul Museum/march/stoneanchors/Stone_Anchors.html
342