27

The first Lapita pottery found in Fiji: Links to an early Pacific world

Matthew Spriggs

Accession 948 in what is now the Phoebe A. Hearst Museum of Anthropology of the University of California Berkeley, logged in on 25 October 1948, does not on the face of it seem such an interesting collection.¹ It consists of an almost random collection of potsherds of different styles and ages, a stone adze and a nondescript stone flake collected from the Sigatoka Sand Dunes on Viti Levu in Fiji. It was sent to Director Edward Gifford at the museum by a medical doctor, Lindsay Verrier, who worked for the colonial administration in Fiji. At the time, Fiji was still ruled by the UK as a Crown Colony, becoming an independent nation in 1970.

Figure 27.1. Location of the Sigatoka Dune Complex.

Note: Version from Dickinson et al. 1998:4.

Source: Courtesy of David Burley.

Figure 27.2. Geomorphic sketch map of the Sigatoka Dune Complex.

The asterisk near the western end of the Dune Complex marks the Nagarai site VL 16/22 (version from Dickinson et al. 1998:5).

Source: Courtesy of David Burley.

We now know that among the collected artefacts were at least three (two were subsequently glued together) dentate-stamped Lapita sherds. Lapita is the name of the style of pottery representing the first settlement, about 3,000/2,800 years ago, of those parts of Island Melanesia and Western Polynesia beyond the main Solomon Islands chain, the area called by archaeologists ‘Remote Oceania’, including Vanuatu, New Caledonia, Fiji, Tonga and Samoa. Its full distribution is from the island of New Guinea, through the Bismarcks and Solomons and east as far as Samoa (Kirch 2017:74–106). Accession 948 included the first sherds of Lapita pottery found in Fiji and a key clue as to its first settlers. But it took some years before its significance was realised, the story we will tease out below (see also Howes, Chapter 15, and Spriggs, Chapter 24, both this volume for more on the story of Lapita).

The Sigatoka Sand Dunes (Figure 27.1) are among the largest and highest in the Pacific Islands, covering 240 ha and attaining a maximum height of nearly 60 m. They are fed by sand brought down the Sigatoka River, which defines their eastern end and whose freshwater outflow has prevented the growth of a protective reef. They are thus exposed to the full force of the waves. Sand blowouts periodically expose ancient land surfaces with cultural remains on them, the earliest dating to well before the dunes themselves began to form about 1,500 years ago, back to near the initial settlement of Fiji about 3000 BP (Figure 27.2). The dunes have been subject to sometimes intensive archaeological interest since the 1940s, with ongoing excavations as burials and other remains come to light after storms. The landscape is continually shifting, covering up and revealing a rich archaeological record (see Anderson et al. 2006; Burley and Connaughton 2010; Burley and Dickinson 2004, and references therein).

On 24 February 1947 Edward Gifford and his wife Delila (Figure 27.3) had stepped ashore in Suva, Fiji, from the MV Thor I to undertake the first major post–World War II archaeological expedition in the Pacific Islands. During a nearly seven-month stay, Gifford surveyed for sites on the largest island of Viti Levu (including among the Sigatoka Dunes) and excavated at two significant places, Navatu and Vuda, directed there by the then head of the Fijian (Native) Administration, Ratu Sir Lala Sukuna (Gifford 1951; see also Spriggs 2019). Sukuna had sent along a young Fijian chief, Ratu Rabici Vuikandavu Logavatu, to assist Gifford and also be the eyes and ears of the Fijian Administration on the project; the two became firm friends and continued to correspond until shortly before Gifford’s death (see Spriggs, Chapter 28, this volume for further information on Logavatu). The excavations produced a nearly 2,000-year-old sequence of occupation for the island when the first radiocarbon dates were published in 1955, only five years after this direct dating technique became available to archaeologists.²

Figure 27.3. Delila and Edward Gifford in San Francisco, immediately prior to their departure for Fiji, 1947, from the frontispiece of Delila’s Fiji scrapbook, in possession of Mrs Maureen Frederickson.

Source: Photograph courtesy of Mrs Maureen Frederickson.

The major monograph on Gifford’s work was produced in a most timely fashion in 1951, printed by the University of California Press (Gifford 1951).³ In it was additional information derived from the small collection sent by Dr Verrier, at the time the acting medical officer for Nadroga-Navosa at Lawaqa, where fortuitously Ratu Rabici Logavatu was also now stationed as a provincial scribe. Verrier and Gifford had met during Gifford’s expedition the previous year, when Verrier was posted as the medical officer at the main airport serving Fiji at Nadi. He had visited Gifford’s excavation at the Vuda site in company with the US Army Representative in Fiji, Captain Leo Moore (University of California, Berkeley [UCB], Bancroft Library, CU-23, Box 187, Leo Moore to Gifford, 22 October 1947 and Gifford to Moore, 12 November 1947). In September 1948 Verrier wrote to Gifford about a package of artefacts from Sigatoka that he was separately sending to him in San Francisco:

at one or two spots there have been appearing weathered skeletons and about a dozen house-platforms or mere fire-hearths with artefacts. I have been down several times and have gathered (and now send you) all samples of interest, bearing decoration, also a broken axe, and a flint that seems to have been used as a scraper. You will see that two of the pieces (3 really, but 2 of them fit neatly into a single whole) have queer decoration in disjointed lines. I cannot imagine how this can be done, nor can anyone I have asked. The people nowadays do not do anything like this, I think. (UCB, Bancroft Library, CU-23, Box 187, Verrier to Gifford, 11 September 1948)

From the photograph included in Gifford’s monograph (reproduced in Spriggs, Chapter 28, this volume, Figure 28.2) it is clear that this ‘queer decoration’ (Figure 27.4) was in fact Lapita-style pottery (Gifford 1951:Plates 19c and d).

In the same letter, Verrier noted ‘I have been down several times (once with Rabici)’ and presented a theory on the development of the Sigatoka Sand Dunes based on human-induced erosion of the uplands of the Sigatoka Valley, a hypothesis largely confirmed by subsequent archaeological and geomorphological research (Anderson et al. 2006; Dickinson et al. 1998).

Figure 27.4. ‘Queer decoration’: The Lapita sherds sent by Lindsay Verrier to Edward Gifford in 1948.

Source: Photograph by Matthew Spriggs.

The package sent by Verrier was eventually received on 25 October 1948, accessioned as No. 948 and catalogued as 11/3267-3291 (UCB, Accession 948, copy of Hearst Museum Accession files). Gifford thanked Verrier the next day and sent him an international money order to cover the postage costs for the package (UCB, Bancroft Library, CU-23, Box 187, Gifford to Verrier, 26 October 1948).⁴ The following day Gifford wrote to Logavatu, noting that ‘you and he [Verrier] were on a collecting trip in the sand dunes’ and seeking to establish where exactly the artefacts were found; in Verrier’s original letter of 11 September he had merely noted that they came from the ‘west side’. Gifford and Logavatu had themselves visited the dunes on 13 May and 15 May 1947 accompanied by Nemani Tubou of Volivoli village and the Mbuli (local chief) of Sigatoka, finding artefacts at two separate sites described by Gifford as ‘Eastern Singatoka Sand Dunes (Site 20)’ and ‘Nanggarai, Western Singatoka Sand Dunes (Site 21)’ (Gifford 1951:251–252).⁵ From Logavatu’s reply of 18 November 1948 Gifford determined that the site the pottery came from was Site 21 and stated so in his monograph. Logavatu had written (phrasing as in original):

All specimens were collected from the same places we hunted but most from the spot where brown soil visible. There also few pieces which you may have noticed have wavy or ?cross markings similar to some of Uluinavatu. Those were found on different location. They are from a valley pass [sic: past] our starting point towards Yadua, in other words on Lautoka side. If I remember well our trip to sand dunes with Dr Verrier was in May, probably on the 16th. (UCB, Bancroft Library, CU-23 Box 187, Rabici to Gifford, 18 November 1948)⁶

The description of a valley past their previous starting point towards Yadua must refer to a location somewhere west of where Gifford and Logavatu had observed and collected material at Site 21, with the valley referred to presumably being the Yalasuna River/Creek valley that debouches immediately east of Yadua village (see map in Kumar et al. 2004:112).

The ‘wavy or ?cross markings’ presumably describe the Lapita sherds, further confirming that they were found on the trip when Logavatu accompanied Verrier. Logavatu’s interpretation of them as similar to that which Gifford and he had excavated at Navatu appears to have influenced Gifford’s later description of them as representing the ‘later Fijian style’. In both cases this was because of unfamiliarity with this ‘new’ type of pottery and the small size of the sample available.⁷ The description of the Lapita sherds as coming from further west than the other sherds collected by Verrier would seem to confirm Gifford’s interpretation of an origin from Site 21 sensu lato. It was – in the absence of knowledge of this key piece of correspondence – later disputed by Green and Palmer (1964). They adjudged the Lapita sherds to have in fact come from Site 20 instead (later renumbered by the Fiji Museum as site VL16/1). Green and Palmer published this contention in Journal of the Polynesian Society, admitting that they had asked Verrier himself to describe where he had found the sherds ‘without informing him of our suspicions’ (1964:329): but which sherds? Logavatu was describing two separate sites, one to the west with Lapita sherds (from his description to the west of where they had previously visited at Site 21) and the other further east along the dunes, which almost certainly was Site 20. Gifford too seems to have failed to appreciate this distinction but was at least more correct in attributing the Lapita sherds to his Site 21 area – that is, to the western end of the dunes – than to Site 20, towards the eastern end.

The Green and Palmer article drew a vigorous response from Verrier in a letter to the editor (1965). He wrote:

The two gentlemen who called on me on 24th August last without, as they report, telling me what they had in mind, succeeded so well in hiding their aim that they managed to conceal their meaning too. Consequently your readers may be assured that any site-attributions made by my friend the late Professor E.W. Gifford remain, for the present, so far as I am concerned, precisely where he left them. (Verrier 1965:125)

Later academics, following Palmer’s original dismissal of Verrier’s ‘retraction’ (Palmer 1966:373), appear simply to have ignored his objections. But in fact the clue to the real location of the first Lapita pottery found at Sigatoka is perhaps to be found on the next page of Palmer’s article, where he notes that ‘other Lapita sherds have come from an excavated site VL16/22 which lies well to the west of Gifford’s site 21’ (Palmer 1966:374). He was here referring to the 1965 excavations of Lawrence and Helen Birks of the site of Nagarai as part of the Fiji Museum’s Sigatoka Research Project – significantly the site name ‘Nanggarai’ that Gifford attributed to Site 21. Palmer had also the previous year admitted in relation to site VL16/1 (Gifford’s Site 20) that the Lapita sherds supposedly from there illustrated by Gifford ‘are the only examples amongst the tens of thousands examined at Sigatoka’ (Palmer 1965:26), although later research did reveal a small number of dentate-stamped vessels (Birks 1973). The lack of such sherds at the time, however, appeared to give him no cause to doubt his and Green’s ‘revision’ of the find spot to the eastern end of the Sigatoka Dunes, and also does not appear to have troubled other archaeologists who have followed their lead!

It could be argued that it would be more plausible to attribute the Lapita sherds found by Verrier and Rabici to present-numbered site VL16/22 Naqarai or its vicinity (arguably included within Gifford’s conception of Site 21) than it is to attribute them to VL16/1 (certainly Gifford’s Site 20) as every archaeologist since Green and Palmer has done. Or, rather, it might be had Birks (1973:113) not illustrated an almost exactly identical Lapita sherd to Gifford’s (1951) Plate 19c, one of five sherds of Pot 55, said to have been excavated from Locality 42A/X within VL16/1 in the 1965 season, featuring Mead motif M30.2 (Mead 1975). Birks notes the similarity to the sherd collected by Verrier and Rabici, noting it as ‘probably from the same vessel’ (Birks 1973:27). Palmer (1966:374) had earlier flagged this connection, noting that the Birks ‘have one particular sherd identical with that figured by Gifford and it could conceivably come from the same dish’. It thus appears that identical sherds almost certainly from the rim of the same vessel were found 17 years apart at VL16/1 and that Rabici’s memory of where the earlier Lapita sherds were from was confused.⁸

There is some more recent evidence to back up the attribution of the Verrier/Rabici sherds to VL16/1 in the form of a very distinctive Lapita pot excavated in 2004 from the site and reported on by Burley and Connaughton (2010). They note that the upper band of decoration is Mead’s motif M12.2 (Mead 1975) but failed to observe that this is exactly the same motif (albeit not from the same pot) as that in Gifford’s (1951) Plate 19d, the other dentate-stamped sherd sent by Verrier. On this basis we have to conclude that the reattribution of the Verrier/Rabici Lapita sherds from Site 21 to Site 20 (the later VL16/1) was justified.

Gifford’s opinion of the date of the sherds we now know to belong to the Lapita tradition was that they were associated with the Late period in Fiji; that is, the last few hundred years (Gifford 1951:236). He detected decorative similarities of this ‘roulette-incised’ ware with pottery reported in 1948 from the Île des Pins (Lenormand 1948) and with pottery found during his and McKern’s 1920–21 expedition to Tonga (McKern 1929:Plate VI). No dates were available for any of this pottery when Gifford was writing, but all this was to change after his next expedition in 1952 to New Caledonia, accompanied by Delila Gifford, Richard Shutler Jr and Mary Elizabeth Shutler (Gifford and Shutler 1956).

This included the first formal excavations at Site 13 on the Foué Peninsula near Koné on the west coast, the site recorded by Gifford as ‘Lapita’. The site had first been recorded in the early 1900s (Piroutet 1909), and its ‘pointillé’ decoration noted on a visit by Fritz Sarasin in 1911 (Sarasin 1917a:119, 1917b:122).⁹ This is what we now know as dentate-stamped decoration, produced by impressing toothed stamps of different sizes and lengths into the wet clay of the pot prior to firing. Jacques Avias (1949) had earlier suggested, on the basis of the published description, that pottery from what would later be known as Site 13 or Lapita would turn out to be similar to that from the Île des Pins and more significantly to that from Watom Island, off the coast of East New Britain in the Bismarck Archipelago discovered by Father Otto Meyer (see Howes, Chapter 15, and Spriggs, Chapter 24, both this volume for further details). And so it was. Gifford (1953) had completed the comparison, bringing in the Verrier collection sherds from Sigatoka and the Tongan pottery as well.

How the name for all of this pottery came to be the ‘Lapita style’ is another story (Spriggs in press), but in part it was to do with Site 13 Lapita being the first directly radiocarbon dated site of this culture, with dates of ‘846 B.C.’ and ‘481 B.C.’ being given (Gifford and Shutler 1956:89–92) – this was of course in the days before recalibration of radiocarbon dates to calendar years, which would push the ages back slightly further. This gave the clue that the Sigatoka sherds were early rather than late, and similarly provided a much greater antiquity than thought for the Tongan pottery as well.

Gifford’s journey to establishing the age and full distribution of the Lapita culture had not started in Fiji but in Tonga in 1920–21, but those few sherds sent in Verrier’s parcel late in 1948 were very quickly to assume a key significance. Links between the design on these sherds and those from the Île des Pins led directly to Gifford’s decision to mount his next Pacific archaeological expedition to New Caledonia and the excavations at Site 13 or Lapita. Recognition of this pottery as being part of an early ‘community of culture’ (Golson 1961:176) linking Polynesia and Melanesia led directly to further work at Sigatoka in the mid-1960s that has continued almost every year to the present, and to further elucidation of the spread of this foundational culture through Near Oceania and into Remote Oceania, and the tracing of its links back into Island Southeast Asia.

Objects highlighted in this chapter were on display at the Phoebe A. Hearst Museum of Anthropology from February to March 2020.

References

Anderson, A., R. Robert, W. Dickinson, G. Clark, D. Burley, A. de Biran, G. Hope and P. Nunn 2006 Times of sand: Sedimentary history and archaeology at the Sigatoka Dunes, Fiji. Geoarchaeology 21(2):131–154. doi.org/10.1002/gea.20094.

Avias, J. 1949 Contribution à l’étude de l’archéologie et de la préhistoire néo-calédoniennes: poteries et industrie lithique (notes préliminaires). Comptes Rendus des Séances de l’Institut Français d’Anthropologie, Paris 59(3):26–28.

Birks, L. 1973 Archaeological excavations at Sigatoka Dune Site, Fiji. Bulletin of the Fiji Museum 1. Suva: Fiji Museum.

Burley, D. and J. Connaughton 2010 Completing the story: A late Lapita dentate stamped pot from Sigatoka, Fiji. Archaeology in Oceania 45(3):144–146. doi.org/10.1002/j.1834-4453.2010.tb00090.x.

Burley, D. and W. Dickinson 2004 Late Lapita occupation and its ceramic assemblage at the Sigatoka Sand Dune site, Fiji and their place in Oceanic prehistory. Archaeology in Oceania 39:12–25. doi.org/10.1002/j.1834-4453.2004.tb00553.x.

Cambridge University Library, G.K. Roth Collection, MS. Add. 8780, Box 7, item 65, ‘Excavations’.

Dickinson, W., D. Burley, P. Nunn, A. Anderson, G. Hope, A. de Biran, C. Burke and S. Mataraba 1998 Geomorphic and archaeological landscapes of the Sigatoka Dune Site, Viti Levu, Fiji: Interdisciplinary investigations. Asian Perspectives 37:1–31.

Gifford, E.W. 1951 Archaeological excavations in Fiji. University of California Anthropological Records 13(3):189–288. Berkeley and Los Angeles: University of California Press.

Gifford, E.W. 1952 A carbon-14 date from Fiji. Journal of the Polynesian Society 61:327.

Gifford, E.W. 1953 L’archéologie Néo-Calédonienne en 1952. Journal de la Société des Océanistes 9:64–70. doi.org/10.3406/jso.1953.1768.

Gifford, E.W. 1955 Six Fijian radiocarbon dates. Journal of the Polynesian Society 64:240.

Gifford, E.W. and R. Shutler Jr 1956 Archaeological excavations in New Caledonia. University of California Anthropological Records 18(1):1–148. Berkeley and Los Angeles: University of California Press.

Golson, J. 1961 Report on New Zealand, Western Polynesia, New Caledonia and Fiji. Asian Perspectives 5:166–180.

Green, R.C. and J.B. Palmer 1964 Fiji sequence: Corrections and additional notes for Sigatoka. Journal of the Polynesian Society 73(3):328–333.

Kirch, P.V. 2017 On the road of the winds: An archaeological history of the Pacific Islands before European contact. Revised and expanded edition. Oakland: University of California Press. doi.org/10.1525/9780520968899.

Kumar, R., P. Nunn and W. Dickinson 2004 The emerging pattern of earliest human settlement in Fiji: Four new Lapita sites on Viti Levu Island. Archaeology in New Zealand 47(2):108–117.

Lenormand, M.H. 1948 Découverte d’une gisement de poteries indigènes à l’Ile des Pins. Etudes Mélanésiennes 3:54–58.

McKern, W.C. 1929 Archaeology of Tonga. Bernice P. Bishop Museum Bulletin 60. Honolulu: Bishop Museum.

Mead, S.M. 1975 The decorative system of the Lapita potters of Sigatoka, Fiji. In S.M. Mead, L. Birks, H. Birks and E. Shaw (eds), The Lapita pottery style of Fiji and its associations, pp. 19–43. The Polynesian Society Memoir 38. Wellington: The Polynesian Society.

Palmer, J.B. 1965 Excavations at Karobo, Viti Levu. New Zealand Archaeological Association Newsletter 8(2):26–34.

Palmer, J.B. 1966 Lapita style potsherds from Fiji. Journal of the Polynesian Society 75(3):373–377.

Piroutet, M. 1909 En Nouvelle-Calédonie. L’Anthropologie 20:605–609.

Sarasin, F. 1917a Neu-Caledonien und die Loyalty-Inseln: Reise-Erinnerungen eines Naturforschers. Basel: Georg & Co.

Sarasin, F. 1917b La Nouvelle-Calédonie et des Iles Loyalty: Souvenirs de voyage d’un naturaliste (traduit de l’Allemand par Jean Roux). Bâle: Georg & Co. doi.org/10.5962/bhl.title.30086.

Spriggs, M. 2019 Covert control? Indigenous agency in Edward Winslow Gifford’s Fijian archaeological expedition of 1947. Journal of Pacific History 65(3):397–416. doi.org/10.1080/00223344.2018.1556089.

Spriggs, M. In press Lapita: History of a name, its terminologies and influences [details of volume not yet available].

UCB (University of California, Berkley), Accession 948, copy of Hearst Museum Accession files.

UCB, Bancroft Library, CU-23 Box 187, Rabici to E. Gifford, 18 November 1948.

UCB, Bancroft Library, CU-23, Box 187, E. Gifford to Dr Verrier, 26 October 1948.

UCB, Bancroft Library, CU-23, Box 187, E. Gifford to J. Hagen, 5 January 1949.

UCB, Bancroft Library, CU-23, Box 187, E. Gifford to Rabici 8 January 1949.

UCB, Bancroft Library, CU-23, Box 187, L. Moore to E. Gifford, 22 October 1947 and E. Gifford to L. Moore, 12 November 1947.

UCB, Bancroft Library, CU-23, Box 187, Dr Verrier to E. Gifford, 11 September 1948.

Verrier, L. 1965 Correspondence. Journal of the Polynesian Society 74(1):125.

---