Discovery of a Lapita Sherd Inland of the Northeast Coast :: The University of the South Pacific Library

Expand document

Collapse document

Discovery of a Lapita Sherd Inland of the Northeast Coast

1 Pages 1-10

1.1 Page 1

1.2 Page 2

1.3 Page 3

1.4 Page 4

1.5 Page 5

1.6 Page 6

1.7 Page 7

1.8 Page 8

1.9 Page 9

1.10 Page 10

2 Pages 11-20

2.1 Page 11

2.2 Page 12

2.3 Page 13

2.4 Page 14

2.5 Page 15

2.6 Page 16

2.7 Page 17

2.8 Page 18

2.9 Page 19

2.10 Page 20

Discovery of a Lapita Sherd Inland of the Northeast Coast

1 Pages 1-10

▲back to top

1.1 Page 1

▲back to top

INSTITUTE OF APPLIED SCIENCES

THE UNIVERSITY OF THE SOUTH PACIFIC

Discovery of a Lapita Sherd Inland

of the Northeast Coast of Viti Levu,

Island, Fiji: Insights and Implications

IAS Technical Report No. 2002/5

Roselyn Kumar

June,2002

1.2 Page 2

▲back to top

Discovery of a Lapita sherd inland of the northeast coast

of Viti Levu island, Fiji: insights and implications

Roselyn Kumar

Institute of Applied Sciences

The University of the South Pacific

Suva

FIJI

Contents

Introduction

The Qaqaruku rock shelter

Discussion

Qara-ni-Oso II compared to Qaqaruku (Ra)

Lapita settlement in and offshore eastern Viti Levu

Larger islands or smaller islands

Conclusion

Acknowledgements

References

1.3 Page 3

▲back to top

Figure captions

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Figure 6.

Location of the area studied within Fiji.

Eastern Viti Levu and areas offshore, showing the locations of

the Lapita sites on Moturiki and Naigani, and the area of Figure

3 in which the Qaqaruku site is located.

Location of the Qaqaruku rock shelter within the area of

northeast Fiji studied.

Drawing of the Lapita sherd found on the floor of the Qaqaruku

rock shelter (drawing by Maxly Viro).

Some of the other decorated sherds found on the floor of the

Qaqaruku rock shelter (drawings by Maxly Viro). The upper

piece is intricately designed and includes both incisions and

applique motifs. The middle piece has a comb-incised motif in

its central part and notching along its rim (at the top of the

drawing) which may indicate that it is also of Lapita age. The

lower piece is an incised rim sherd which may also be of Lapita

age although it is too small to show enough of its decoration.

The location of the Qara-i-Oso II (Qwara-i-Oho II) site in

southwest Viti Levu, where a Lapita sherd was discovered by

Anderson et al. (2000).

1.4 Page 4

▲back to top

Plate captions

Plate 1.

Plate 2.

Plate 3.

Plate 4.

The inland village of Naivoco. The rock platform on which the

village has been built is formed from limestone full of fossil

corals and seashells which forms the base of the Tova Andesite

above. The Qaqaruku rock shelter is cut in this limestone and is

off to the right of the photo, below the level of the village.

View seawards from Naivoco Village. The edge of the coastal

plain is here occupied by the villages of Delaiyadua,

Matainananu and Dogoloa. In Lapita times, the sea level was

about 1.5 metres higher than today and the coast may have

been 1-2 kilometres farther inland.

The entrance to the Qara-i-Oso II (Qwara-i-Oho II) cave near

Tau Village in southwest Viti Levu, where a Lapita sherd was

discovered during excavation by Anderson et al. (2000).

View seawards from just below the Qara-i-Oso II (Qwara-i-Oho

II) cave site. In Lapita times, the coast may have be~n 1-2

kilometres farther inland.

1.5 Page 5

▲back to top

Introduction

A geological reconnaissance of the Tova Andesite, a volcanic edifice of Plio-

Pleistocene age, in the province of Ra in the northeastern part of Viti Levu

island (Figure 1) led to discoveries of fragments of earthenware pottery,

including a sherd, unmistakably Lapita in age (2600-2900 cal years BP or

650-950 BC). Discovering Lapita sherds in any particular locality in Fiji is

significant because these are indicative of the first human inhabitants of these

islands. Some of this ancient earthenware is decorated with characteristic

dentate stamps, a style accepted as being unique to the Lapita people (Kirch,

1997).

The Qaqaruku rock shelter

The Tova Andesite in northeast Viti Levu island, Fiji, is the youngest extrusive

igneous formation on the island. Some of the reports of the geologists who

first mapped it in the 1950s stated that the base of the Tova Andesite was in

places marked by a shell bed and fossil coral reef, which was of interest to the

author because of its potential for revealing information about Quaternary sea-

level changes. One place where the shell bed and reef were reported to

outcrop was at Naivoco Village, and one of the best locations found there was

in the walls of a rockshelter named Qaqaruku. After examining the walls of

this shelter, a cursory examination of its floor revealed pottery, and a small

collection was made. This was later found to include a Lapita sherd.

Two Lapita-age settlements are already known from the islands offshore of

eastern Viti Levu (Figure 2). These are the Matanamuani site on Naigani

Island (Best, 1981) and the Saulevu site on Moturiki Island (Nunn, 1999a).

Unpublished temper analysis of Naigani pottery by William Dickinson

suggests that this may have been manufactured and imported from Moturiki,

thereby adding weight to the suggestion that the Lapita occupation of these

islands was connected . This idea has already been proposed as a result of

both geography (Nunn, 1999a) and oral histories (Ramoli and Nunn, 2001 ). It

may be significant, as discussed below, that Naigani Island can be seen from

the Qaqaruku area.

The rock shelter named Qaqaruku lies 250 metres to the west of the inland

village called Naivoco, in the province of Ra, in the northeast part of Viti Levu

island (Figure 3). Fossil corals and seashells in situ cha,racterize the limestone

in which this shelter has been cut (Plate 1). This limestone is an important

stratigraphic marker horizon in the area because it lies at the base of the Tova

Andesite, the youngest suite of volcanic ro,cks on Viti Levu, which erupted

from the late Pliocene (about 2 million years ago) to the middle Pleistocene

(around 1.2 million years ago).

The Qaqaruku rock shelter is not a particularly impressive feature mainly

because it is not presently wide enough or extensive enoug_h to be habitable

by a significantly large human population. The overhang is not broad enough

1.6 Page 6

▲back to top

to provide adequate shelter although it may have been different 3000 years

ago.

In contrast to the grassland-covered hills around, the Qaqaruku site is covered

with lush vegetation and there is a dependable freshwater supply from the

river 200 metres downslope. The river hosts communities of freshwater

shellfish which, judging from the numbers scattered around the base of the

Qaqaruku rock shelter, formed an important part of the diet of the people that

lived there. There is also a rich supply of marine resources three kilometres

away on the ocean shore where the villages of Delaiyadua, Matainananu and

Dogoloa are now located (Plate 2).

A Lapita sherd about 5 centimetres in diameter, exhibiting three parallel rows

of dentate stamps, was found on the floor of the rock shelter. The design motif

covers almost half of the sherd (Figure 4). Three of the other pieces collected

from Qaqaruku are shown in Figure 5. Two of them may also be decorated in

a characteristic Lapita style.

The surface collection at Qaqaruku was made in less than one hour, and

since it included 1-3 Lapita sherds, it is certainly an important site which

deserves to be revisited and systematically investigated.

Discussion

Lapita sherds have been discovered on many coastal plains, paleo-beach

ridges, and areas inland . on oceanic islands throughout the western tropical

Pacific (Kirch, 1997). Discovering Lapita sherds inland on larger islands is not

as common as findings these sherds on smaller islands.

The recovery of a Lapita sherd inland on the large island of Viti Levu allows

new insights into the understanding of the settlement patterns of the Lapita

people. For a number of years it has been widely thought that the Lapita

people favoured smaller islands offshore from larger islands for settlement

(Green, 1979; Kirch , 1997). This view has been challenged by Spriggs (1984)

and others who consider that the concentration of Lapita sites triscovered on

smaller offshore islands is a function of post-settlement geomorphological

processes which modify the coasts of larger islands and thereby obscure

Lapita sites more than they do those of smaller islands. With the discovery of

a second inland site on Viti Levu, combined with the possibility (discussed by

Kumar, 2002) that charcoal production 40 km up the Sigatoka Valley may

have been a result of burning by the Lapita people, it now seems likely that

Lapita settlement was widespread on larger islands, both inland and coastal.

It is reasonable to suppose that one reason why Lapita people occupied the

northeast coast of Viti Levu was because of the abundance of raw materials

for pottery manufacture (Geraghty, 1996). This suggests a prolonged rather

than a short period of occupation.

1.7 Page 7

▲back to top

Qara-ni-Oso II compared to Qaqaruku (Ra)

The inland Lapita site in Ra can be compared to another interesting inland site

near Nadi on the southwestern coast of Viti Levu (Figure 6). An excavation in

a large limestone cave, 250 metres above sea level in a cliff overlooking the

· coast, and about 4 km inland from the nearest present coastline, led to the

unexpected discovery of a Lapita sherd (Anderson et al., 2000) . This Lapita

site, Qara-i-Oso II (actually Qwara-i-Oho II in the language of the local people)

near Tau village neither lacks freshwater supply nor marine resources though

the coast is slightly farther away from the sea than is the situation in Ra. But

the shore near Qara-i-Oso 11 could have been 1-2 kilometres inland of its

present position some 4000 years ago1 at the time of the Holocene sea-level

maximum when the ocean surface in Fiji was at least 1.5 metres higher than

today (Nunn and Peltier, 2001 ).

The landscape around Tau village is dominated by ephemeral rivers and

streams cutting the volcanic bedrock into hills and deep valleys. Stream

channels dry up easily on this leeward side of the island and get replenished

only during times of heavy rainfall.

However, the freshwater supply at Qara-i-Oso ·11 during the Lapita era (2600-

2900 BP) would probably have been greater than today because the climate

is known to have been wetter (Nunn, 1999b). The discovery of a Lapita rim

sherd at the Qara-i-Oso II inland site anticipated the discovery of other inland

sites on the two main islands of Fiji (Anderson et al., 2000) and this has

proven to be the case.

Compared to the Qara-i-Oso II site, the Ra inland Lapita site is about 200

metres higher than the river draining out to the sea and is approximately 3

kilometres inland from the present coast at Ra. A site like this raises several

questions, particularly concerning the reasoning behind the occupation of the

area by Lapita people, which is contrary to the belief that they settled almost

exclusively on small islands (Green, 1979; Kirch, 1997).

Lapita settlement in and offshore eastern Viti Levu

The Ra site lies in proximity to another small coastal Lapita site called

Matanamuani on Naigani island (Best, 1981). The Matanamuani site at

Naigani lies northwest of another Lapita site called Saulevu on Moturiki Island

(Nunn, 1999a). It has been speculated that, since a Lapita settlement existed

at Saulevu on Moturiki, then it is probable given the nature of the offshore

reefs and the barriers they pose to sailing vessels that the inhabitants came

from the northwest, where Naigani lies (Nunn, 1999a). Affiliations between

pottery tempers can also be expected between these two small coastal

islands and the inland Ra site which is nearby (see Figure 2). It is also

probable that the same people who settled in these areas also settled the

1 Note that the Holocene sea-level maximum was about 4000 years ago while the first trace of

Lapita settlement in Fiji was around 3000 years ago, a time when the sea level was still close

to its Holocene maximum level.

1.8 Page 8

▲back to top

leeward coast of larger, well-watered, fertile Ovalau island even though Lapita

potsherds and Lapita sites have not yet been recorded there.

The temper sands in the Lapita sherds found on Naigani were not made using

the local non-placer litho-feldspathic sands (the only non-calcareous sand on

the island) derived from its dacitic outcrop. As the geology of Moturiki is

similar to the geology of Ovalau (Ibbotson, 1961; Coulson, 1976), Dickinson

(2001) concluded that the exotic Naigani temper might reflect ceramic transfer

from either island. However, there is another possibility that the exotic

Naigani tempers might reflect ceramic transfers not only from Ovalau and

Moturiki but also from Ra.

With no petrographic analysis of the temper sand in the sole Lapita sherd from

Ra, there is no conclusive evidence of ceramic (or temper-sand) transfer to

Naigani from Ra, which highlights the need for petrographic analysis of pottery

from the latter area. The examination of the Ra Lapita sherd under a

compound microscope revealed dominantly terrigenous temper sands

believed to have a local provenance, reflecting the local augite-andesitic

lithology of the Tova Andesite. The temper sand in the Moturiki Lapita sherd

is pyribole2-rich dominated by augite and _hornblende (Dickinson, 2001 ). It

resembles the composition of the exotic Naigani temper sand (Dickinson,

2001) . The augite-rich and dominant phenocrysts of the Tova Andesite raise

the possibility that ceramic temper sands used in the manufacturing of the

Lapita sherds at Matanamuani on Naigani were from Ra.

Perhaps the initial colonisers of Naigani arrived from the inland sheltered

areas of Viti Levu. It is plausible to suppose that the initial colonisation route

was from Viti Levu to Ovalau and the small island of Moturiki by way of

Naigani - the island acting as a stepping stone for some of the travellers .

However it is important to consider the proximity of these islands to the

second largest island of Fiji - Vanua Levu - particularly in the tracing of

ancient colonisation routes of the Lapita people. No Lapita sites have yet

been confirmed in southern Vanua Levu and it is likely that the pattern of

Lapita colonisation which is emerging in eastern Viti Levu and offshore is

connected to that which almost certainly occurred in southern Vanua Levu.

This underscores the urgency for exploration on this, the second largest island

in Fiji.

Larger islands or smaller islands

The pattern of known Lapita colonisation in the Fiji Islands suggests that these

people favoured smaller offshore islands. However it is believed, on the

strength of the discovery of a second inland site on Viti Levu, that this view is

incorrect. It is suggested that in fact the Lapita people favoured the larger,

more resource-rich and resource-diverse islands in Fiji, and that it is only the

"visibility" of smaller island sites which has given rise to the present

distribution of known sites. This idea is in line with that of Spriggs (1984).

2 Pyribole : temper sands dominated by augite minerals followed in slightly lower amounts by

hornblende. It is used to calculate the pyribole index defined by Dickinson (200 l ).

1.9 Page 9

▲back to top

If it is assumed that the Lapita voyagers occupied the large islands such as

Viti Levu first and stayed there for several generations, then perhaps the

smaller islands distant from the main islands were deliberately saved for later

exploitation and colonisation . If this is the case, then larger-island sites may

be older than those on smaller islands (from which most settlement ages

come). In this context, the recent re-dating of the Yanuca (Viti Levu) site to

around 3000 BP or earlier (Clark and Anderson, 2001) is significant.

Another aspect of Lapita colonisation in Fiji has also become apparent

recently. The Lapita people seem to have chosen to settle within sight of

another island which was or became settled. Examples include Vanuabalavu

and Mago in the northern Lau islands, Moturiki and Naigani in central Fiji ,

Beqa and Ugaga, and Vatulele and the south-coast sites of Viti Levu. It may

be that intervisibility was an important criterion in the selection of new

settlement sites . One possible reason for this is because seafaring between

islands was so dangerous that it was important to keep vessels plying

common routes between islands within sight. Perhaps it was also so that a

main-island settlement could control access to the sparsely inhabited

surrounding islands, which would have been minimally exploited and been

kept aside as reserved pools of (marine and terrestrial) resources.

The Lapita people may have landed and colonised the larger islands first and

later set out to sail to those smaller intervisible uninhabited islands which were

habitable and large enough to support minimum population numbers .

Anderson, in favouring a late settlement model for East Polynesia, proposed

that " predation took early precedence over horticultural expansion and that

since preferred faunal resources were depleted fairly easily and quickly, there

was an incentive for some people to move on in search of new reserves quite

soon after colonisation instead of remaining to invest more energy in

horticultural expansion" (Anderson 1996: 367). Anderson proposed that

faunal predation resulted in the depletion of the most preferred resources

easily and quickly. On a large island this could have happened when the rate

of consumption of resources was higher than their rate of renewal. To avoid

food shortages it is plausible to suppose that the Lapita people populated the

rest of the nearby islands in order to disperse a large population from the main

islands and have sufficient resources per population group.

The small islands, peripheral to the main islands and within sight of another

island, may have also been favoured for settlement because these islands

offered ancient oceanic dwellers easier access (their movements on large

islands may have been restricted by the rugged terrain) to the inland areas,

which were all closer to the coast, and to the coast itself for higher quality

marine foraging, an outcome of the lack of large rivers smothering nearshore

reefs with sediments.

1.10 Page 10

▲back to top

Conclusion

This paper does not completely abandon the popular view that Lapita people

settled mostly on smaller islands within the archipelagoes of the western

Pacific but provides new evidence that they also colonised larger islands in

presumably sizable numbers. It is clear that this colonisation was both inland

and coastal.

Lapita settlement of larger islands has been more difficult to detect than that

on smaller islands. Yet it is possible that Lapita settlements on larger islands

in archipelagoes like Fiji were more numerous, more extensive in the area

they covered, and even older than the Lapita settlements on the smaller

islands.

To fully comprehend the history of the prehistoric Lapita settlements,

researchers need to move their exploration inland on larger islands

particularly to habitable sites proximal to the coast where freshwater is readily

available, and undoubtedly more findings of inland sites will be made. Further

in-depth study or field research may bring forth additional evidence for inland

Lapita colonisation on Viti Levu and may also reveal this on Vanua Levu.

Acknowledgements

I am grateful to Bill Aalbersberg (Head of the Institute of Applied Sciences at

the University of the South Pacific) and Patrick Nunn (Head of the Geography

Department at the University of the South Pacific) for their critical editing of

this paper, and to the people of Naivoco and Delaiyadua villages in Ra.

References

Anderson, A., Clark, G. and Worthy, T. 2000. An inland Lapita site in Fiji.

Journal of the Polynesian Society, 109, 311-316.

Best, S. 1981. Excavations at Site VL 21/5 Naigani Island, Fiji: a preliminary

report. Department of Anthropology, University of Auckland.

Clark, G.R. 1999. Post-Lapita Fiji: cultural transformation in the mid-sequence.

Unpublished PhD thesis, The Australian National University, Canberra.

Clark, G.R. and Anderson, A. 2001 . The age of the Yanuca Lapita site, Viti Levu,

Fiji. New Zealand Journal of Archaeojogy, 22, 15-30.

Coulson, F.1.E., 1976. Geology of the Lomaiviti and Moala island groups. Fiji

Mineral Resources Division, Bulletin 2.

Dickinson, W.R. 2001. Petrography of temper sand in a Lapita sherd from

Moturiki in Fiji. Unpublished petrographic report WRD-205, 3 pp.

2 Pages 11-20

▲back to top

2.1 Page 11

▲back to top

Geraghty, P. 1996. Pottery in Fiji: a preliminary survey of locations and

terminology. In: Davidson, J.M., Irwin, G., Leach, B.F., Pawley, A. and

Brown, D. (eds.). Oceanic Culture History: Essays in Honour of Roger

Green. New Zealand Journal of Archaeology Special Publication, 421-

431.

Green, R.C. 1979. Lapita. In: Jennings, J. (ed.). The Prehistory of Polynesia.

Cambridge: Harvard University Press, 27-60.

Ibbotson, P., 1961. The geology of Ovalau, Moturiki and Naingani. Geological

Survey of Fiji, Bulletin 9.

Kirch, P.V. 1997. The Lapita Peoples. Oxford: Blackwell.

Nunn, P.O. 1999a. Lapita pottery from Moturiki Island, central Fiji. Archaeology

in New Zealand, 42, 309-313.

Nunn, P.O. 1999b. Environmental Change in the Pacific Basin: chronologies,

causes, consequences. London: Wiley, 357 pp.

Nunn, P.O. and Peltier, W.R. 2001. Far-field test of the ICE-4G (VM2) model of

global isostatic response to deglaciation: empirical and theoretical

Holocene sea-level reconstructions for the Fiji Islands, Southwest

Pacific. Quaternary Research, 55, 203-214.

Ramoli, A. and Nunn, P.O. 2001. Naigani Island and its historical connections

with Ovalau and Moturiki Islands: convergences between legend and

fact. Domodomo, 13, 19-28.

Spriggs, M. 1984. The Lapita Cultural Complex: origins, distribution,

contemporaries and successors. Journal of Pacific History, 19, 202-

223.

2.2 Page 12

▲back to top

178°E

180°

-17°s

•

Waya-- q Natu7uku

~ Sigatoka. ·•ilil!i!liil!lill l l liiil ili!ll!!il!ili!lit

p\\

::::::::::·;:::·:::::::;::::.::· ::::::::::·:·.

'\\)

dunes

0 D>

Vatulele - - ~ se{a

-1s0 s

50 km

location of

•!

Figure 2

FIGURE 1

Vanuabalavu

)

Mago CJ

r.,)

Lakeba@

if"

(}

Totoya

2.3 Page 13

▲back to top

<,/J 178030'E

: .~

·,o

°'

Area of

Figure 3

17°45'. S

Naivoco_ ·

Lapita site

· (Qaqaruku)

...v.·· ·I.·t·I Le.·v.a.

178°45'E

FIGURE 2 .

2.4 Page 14

▲back to top

·- ; Nayavutoka

Dogoloa

mangrove

· swamp

Matainananu

Delaiyadua

fringing

reef edge

/ \\ ..

Nasinu

Qaqaruku

rock shelter

FIGURE 3

481

Tova Peak

A.646

1 - ~ ,I

2.5 Page 15

▲back to top

LAP,rA

~ "j 01 rvkv

- - -- - - - - - - -.2f!':!:".::>-...-,

• •~ e o .

eti,ell · ,, _

-C, - ~ - - ,,z_,

FIGURE 4

2.6 Page 16

▲back to top

- - -- -- ---

FIGURE 5

2.7 Page 17

▲back to top

Momi

Bay

! _,.,,..-

•·... ,_,·

fringing

reef edge

FIGURE 6

2.8 Page 18

▲back to top

~,.,-\\ ,..._,,

2.9 Page 19

▲back to top

;1·

!'.--/,

,:]"-- .

'•

"tt;'.:;!~:!'• :;.,,,; .--··

•·::-:: ·

•-:-:,~ '

; '!•

.J~ii-~'.:~.~a(li;,lb.

'"'·

.. . 4 ,./

2.10 Page 20

▲back to top

:i ~- !.~;/1:-,:. ·••'-', .··