Kava Bowls From Fiji, Samoa, and Tonga
KAVA BOWLS FROM FIJI, SAMOA AND TONGA
© Valentin Boissonnas 2021
Unlike figurative arts from Melanesia and Polynesia, utilitarian artefacts have often languished in the shadows of research and been neglected by collectors. That has to be deplored because often such objects were intricately linked to the lives and beliefs of their owners and can document complex patterns of exchange and human interaction. Kava bowls from Western Polynesia are an excellent example which I’d like to explore in this paper as their shapes and associated rituals give us precious insights on how the use of kava evolved in these island societies. This discussion is based on the more detailed and referenced Beyond the rim: a comparative study of kava bowls from Samoa, Tonga and Fiji that can be found in the Journal of the Polynesian Society Vol. 123, no. 4, 2014, pp. 357-382. For the study more than a hundred early bowls with recorded provenances were studied. All bowls with a MAA object number have been reproduced by kind permission of University of Cambridge Museum of Archaeology & Anthropology.
Kava, also known as qona, aqona or yaqona in Fiji, is an infusion of gray-brown colour made from the root of a pepper plant (Piper methysticum) with an acquired earthy and woody taste. Its effects can be attributed to its active ingredient, kavalactones, which affects the Central Nervous System causing muscle-relaxation, stress reduction and clear-headedness. Its use in Polynesia was most likely introduced from Vanuatu via Viti Levu. In Fiji yaqona was reserved for Melanesion-derived būrau practices during which a priest (bete) would invoke the presence of a god. For that yaqona was grated (unlike in Polynesia where it was masticated), mixed with water in a bowl, filtered through a wooden or wickerwork filter and poured into a shallow drinking cup or dish. The liquid was then sucked from centre of the dish. Judging from the first-hand account of Methodist Rev. Thomas Williams (Williams 1858: 225), who lived and worked in the middle of the 19th century in Lau, Taveuni and Vanua Levu, yaqona was consumed by the bete at the end of the būrau rite as an offering to the departing god inside him. Given the ecstatic state he had worked himself into, yaqona must have been a most welcome drink at the end of the ceremony. Yaqona was also prepared and sucked from circular earth pits lined with vudi plantain or giant taro leaves, possibly a precursor to the clay and wooden bowls (Clunie 1986: 169, 1996: 8; Lester 1941: 111-12).
We do not know when kava consumption reached Sāmoa, but myths indicate that it was introduced to the two islands through Fijian contacts. In Sāmoa kava drinking changed into a highly structured and ritualized communal drinking that reflected social tissue, rank and status. Ceremonial specialists (tūlāfale) orchestrated the ceremonies according to a very specific protocol while attendants were seated in a circle, reflecting Sāmoan society. It is likely that with the introduction of kava in Sāmoa it was transformed from a single-person sacred būrau rite into a chiefly ceremony with the participation of a multitude of high-ranking individuals.
Tongan tradition relates the introduction of kava to the reign of the 10th Tu‘i Tonga, therefore approximately to the 12th or 13th century (Gifford 1929: 156). The organisation of the Tongan kava ceremony is a clear indication that it was adopted from Sāmoa. The ritualised and formal part of the ceremony continued to be handled by ceremonial specialists of Sāmoan descent (matāpule). As outsiders and worshippers of their own “foreign” gods, they were not bound by local taboos and thus physical contact with high-ranking chiefs was not an issue.
In the mid-18th century two master carver clans, originating from Manono Island in Sāmoa, were resettled under the patronage of the Tu‘i Tonga on the island of Kabara (Lau archipelago) where the best and largest vesi (Intsia bijuga) groves stood (Clunie 2013: 180). This highly desirable and resistant hardwood was not only ideal for house and canoe construction, but also the preferred wood for war clubs, priestly oil dishes and kava bowls. After the premature death of Lehā, the Tu‘i Tonga’s principle carpenter and canoe builder, his junior kinsman Lemaki and his descendants became the dominant canoe builders and kava bowl producers on Kabara. The introduction of the Micronesian rigged Tongan/Fijian sailing canoes such as the kalia/drua or the hamatafua/camakau can also be attributed to the Lemaki carvers. These particular vessels were built as a replacement for the older clumsier sailing canoes such as the Polynesian-rigged tongiaki. This technological revolution allowed the Tongans to become the dominant trading power in Eastern Polynesia, contributing not only to the exchange and redistribution of valuables such as red feathers, mats, pottery, weapons, head rests, coconut oil and sandalwood, but also kava bowls. It is through this Tongan dominated trade network that the Sāmoan-Tongan derived kava ceremony was finally brought to Fijian shores around AD 1000-1200, thus closing the cultural journey of this extraordinary plant and its associated rituals. In Fiji this new and more social kava drinking coexisted with the original sacred būrau ceremonies until the latter became obsolete with the growing influence of missionaries in the 19th century. In the following chapters an attempt is made to show how bowl shapes can inform us on their makers.
Fijian yaqona bowls
Until the introduction of the Sāmoan/Tongan kava circle in eastern and north-eastern Fiji the consumption of yaqona was reserved for priests and chiefs who consumed it as part of indigenous būrau rites. On such occasions small quantities of yaqona were required and therefore the bowls produced were of diminutive size. The earliest bowl type is a circular bowl named dariniyaqona or sedreniyaqona. Glazed ceramic ones (Fig. 1 left) appear in the archaeological record from AD 1500 onwards (Marshall et al. 2000: 92). Their wooden homologues (Fig. 1 right) have a diameter of 25-35cm and are frequently decorated around the rim. When not in use dariniyaqona are hung from a coir sennit suspension cord that is either passed through two rim perforations or a lateral pierced suspension lug, a feature absent in the dari bowl used for domestic and cosmetic purposes. The semi-circular base of dariniyaqona requires to be stabilised by the use of a plaited or wooden ring (toqi) (see Heerle & Carreau 2013, p. 41, fig. 3.33).
Figure 1: Left: A ceramic dariniyaqona or sedreniyaqona with coir sennit and notched rim collected by Arthur Gordon in the 1870s (British Museum Oc, Fi.12). Right: a wooden example with four raised double lines on the rim area collected by Captain R. E. Stewart in 1877 (MAA 1937.322).
A separate class of bowls are daveniyaqona or ibuburau dishes that can have circular, humanoid or bird-shaped forms and sit on an elaborately carved stand. They are intricately linked to the būrau way of yaqona consumption. These highly sacred pieces only briefly appeared on the market after the mid-19th century when traditional belief systems eroded as a direct result of missionary presence. Discarded as pagan paraphernalia an importantnumber was collected in the 1870ies by collectors such as Sir Arthur Gordon and Anatol von Hügel.
Figure 2: Three daveniyaqona carved from vesi wood. The first one (MAA Z3373) was collected by Arthur Gordon most likely in 1875, the other two (MAA Z3351, MAA Z3364) by Anatole von Hügel in 1876.
Shallow oval or lenticular bowls with pointed ends, much used in Giji and Lau, are generally under 30cm long, and called draunibaka (leaf-of-baka tree). They refer to the baka (Ficus obliqua) tree, which was considered sacred by Fijians as ancestor spirits inhabited them (Parham 1972: 138). Draunibaka often have four stubby sucu feet, some three-legged ones can have a handle as illustrated by Lester (1941: 97, pl. IIB). Legless examples are sometimes referred to as bavelo (dugout or canoe without outrigger). Some deeper draunibaka often lack legs and allow mixing of the yaqona in the bowl. The liquid is then drunk from small coconut cups (bilo), an innovation that became popular with the Tongan kava circle in the early 16th century.
Figure 3: A lenticular draunibaka (MAA Z3475) collected by Anatole von Hügel in 1875.
Larger circular or slightly lenticular four-legged bowls with two pointed sides clearly distinguish themselves from draunibaka by their size and allow mixing of the yaqona in the bowl (Fig. 4). Provenanced specimens were collected in Nadrogā in southwest Vitilevu, Bau in southeast Vitilevu, and in the Lōmaiviti group. The length of those studied generally ranges from 30 to 50cm, their width from 20 to 36cm. Their underside is frequently decorated with two ridges that start from the pointed rim and taper off towards the centre. If botanical at all, the origin of the shape of these bowls could be the seed pod of the tropical almond tavola (Terminalia catappa) that can be found in the littoral and lowland forests of both Melanesia and Polynesia.
Figure 4: An almost circular yaqona bowl (MAA Z3492) with the outside ridge tapering off towards the centre.
Only few bowls have been collected in the western highlands of Vitilevu. They have a deep circular bowl, four elongated legs and a diameter ranging from 25 to 35 cm (Fig. 5). The bowls are well finished and their rim decoration can be notched like on ceramic and wooden dariniyaqona. The legs however can look surprisingly clumsy and do not seem to be part of a well-established canon. It is possible that they represent an early type of four-legged bowls that might have evolved out of wooden dariniyaqona. Given the likely presence of Sāmoan mātaisau in the region in the 16th century (Clunie 2013: 164) they could represent a marriage of legless dariniyaqona with four-legged early Sāmoan ‘ava bowls. The use of stone tools as well as heavy patination from generations of handling and use testify to the great age of some of these archaic bowls.
Figure 5: A tānoatavatava-type bowl (MAA Z3421) collected in the western highlands of Vitilevu by Alfred Maudslay in 1875.
Circular bowls with shorter legs and a similar or larger diameter were also been collected in coastal areas, although their exact origin is not known (Fig. 6). Unlike the highland bowls of western Vitilevu, they are shallower, have thinner walls and have more diversified lug and rim shapes. By the 1900s these bowls were called tānoatavatava to distinguish them from their Tongan-derived lipped counterpart, the tānoa discussed in the next chapter. Tavatava denotes a simple upwardly pointing rim.
Figure 6: A large tānoatavatava (MAA Z30106) with notched rim decoration collected by Anatole von Hügel in 1875.
Turtle-shaped yaqona bowls were comparatively common on Vitilevu, particularly along the north-eastern coast or Rā. The depiction of a turtle associates these bowls to the zoomorphic daveniyaqona dishes (Clunie 1986: 175). The addition of four or more legs on some of them seems to be a later phenomenon, the early pieces all being legless in the Fijian dariniyaqona tradition. A paramount example was collected by James Calvert in 1886 (Fig. 7). Both the large size (Ø 97cm) and the tānoa style rim suggest that it is of Lauan origin and quite possibly from Kabara. The carving is simple and there is no evidence of yaqona use. The popularity of turtle bowls already in the 19th century is illustrated by a four-legged example that Augustin Krämer collected in 1895 in Apia, Sāmoa (Krämer 1994, Vol. II: 245, Fig. 73). With growing tourism turtle-shaped bowls became increasingly popular and smaller sized ones are still being made for sale today.
Figure 7: Left: A turtle-shaped yaqona bowl (MAA 1937.321) given by Rātū Seru Cakobau to Captain R. W. Stewart R.E. in c. 1876. Right the large four-legged turtle-shaped tānoa (MAA Z3972) collected by James Calvert, probably in 1886, and subsequently in the collections of W. D. Webster and von Hügel.
The study of Fijian yaqona bowl profiles clearly shows that Fijian bowl types have a rim that is defined by the thickness of the wall as illustrated in fig. 8. Even though the rim area can be decorated by adding notches or, as found on some examples, by an additional raised band below the outer rim area, it is essentially directed upwards.
Figure 8: Fijian rim profiles found on dariniyaqona, daveniyaqona, draunibaka and tānoatavatava.
Tongan kumete kava and tāno‘a
In Tonga both circular and lenticular kava bowls are generally referred to as kumete kava, the bowl used by the Tu‘i Tonga however was named tāno‘a (Gifford 1929: 161). Like elsewhere, fehi (Intsia bijuga) was the most sought-after hardwood for kumete kava. According to Whistler both feta’u (Calophyllum inophyllum) and tamanu (Calophyllum neo-ebudicum) were also being used for making kava bowls. Ngesi (Manilkara dissecta), kau (Burckella richii), manaui (Garuga floribunda) and mo’ota (Dysoxylum forsteri) were other wood species out of which kumete for food preparation and presentation were fashioned (Whistler 1991: 31-119).
Documented Tongan kumete kava are extremely rare. The only eight existing provenanced circular bowls were collected during the voyages of Captain James Cook, Alejandro Malaspina and Dumont d’Urville. They have diameters ranging from 37 to 72cm and their height ranges between 11 and 17cm. Unlike their Sāmoan and Fijian counterparts the rim area of Tongan bowls collected in the 18th century opens up in a unique outward motion (Fig. 9).
The earliest two examples were collected by Johann and Georg Forster in 1773-74 and clearly show this tentative to open up the rim area (Figs. 9a & b). The bowl collected by Midshipman James Ward on Cook’s third voyage in 1777 amplifies this movement giving the rim a curved wavelike shape (Fig. 9d). Curved rims can also be found on bowls collected by d’Urville 50 years later (Figs. 9f, g and h).
Figure 9: Rim profiles from kumete kava collected in Tonga between 1774 and 1827.
Cook describes a very large bowl from which he was served kava in a plantain leaf cup (pelu) at Mu‘a in 1777 during the mourning ritual for one of the sons of Tu‘i Tonga Fatefehi Paulaho (Beaglehole 1967: 141). The bowl held four to five gallons of liquid, the equivalent of around 20 litres. Given the size and occasion it might very well have been the Tu‘i Tonga’s tāno’a. During his stay in Tonga between 1806 and 1810, William Mariner equally witnessed the use of large bowls during important ceremonies with diameters of up to 91cm and depths of 30cm (Martin 1827, v II: 156). Such exceedingly big bowls could not have been produced in Tonga because of the lack of suitable big fehi trees. They were the product of Lauan workshops on the island of Kabara.
Two lenticular kumete were collected in Tonga by Cook. One is in the Weltmuseum Wien, another was in the George Ortiz collection. Their rims differ from Fijian or Sāmoan counterparts by having both the inner and outer walls of the bowl meeting in a pointed tip, rather than the inside wall ending in a rounded ellipse. Labillardière (1800, Plate 31), illustrates a lenticular kumete with an elliptical Fijian type rim. Even though collected in Tonga, the rim type could indicate a Fijian origin.
Lauan tānoa
As already mentioned the Lau group was home of the prolific Lemaki carvers who produced not only ships, clubs and headrests, but also innumerable of kava bowls. From Lau these bowls were exported to Fiji, Tonga and (via Tonga) to Sāmoa by Tongan navigators. The large size of many of them (their diameter varies between 35 and 100cm) and the particular treatment of the rim area help us to identify them. Unlike the previously discussed Fijian tānoatavatava, the Lauan tānoa rim extends either horizontally or in a gentle curve. It is possible that tānoa profiles derive from the tāno’a, that originally was Tu‘i Tonga’s prerogative. With the waning influence of the Tu‘i Tonga the tānoa type could have become less tapu and more accessible to other chiefs. Both rim types can be seen as stylistic continuations of Tongan bowls collected in the late 18th century (Fig. 10).
Figure 10: Two characteristic types of tānoa profiles encountered in the survey
Some very large tānoa can have six or more legs. In Fiji the arrival of tānoa bowls was immortalized by the naming of Tānoa, future Vūnivalu of Bau, who died in 1852 (Clunie 1986: 173). It is therefore likely that the tānoa was introduced to Viti Levu in the late 1700s. From early travel accounts we know that in Vitilevu and Vanualevu tānoa remained a rare commodity throughout the 19th century. The variations in bowl cavity and rim profiles could indicate that several production centres existed in Lau or that some were carved elsewhere.
Figure 11: Three tānoa all collected in Fiji showing typological variations that are most likely the result of different workshops: The first two (MAA Z3973, MAA Z3984), collected by Sir Arthur H. Gordon, have a curved extended rim but show differences in height, lug and leg shape. The third one (MAA Z30939) was collected by Walter Coote before 1882 and has a horizontally extended rim.
Sāmoan ‘umete and tānoa ‘ava
Sāmoan kava bowls were made from a variety of hardwoods and can be divided into oval or lenticular ‘umete and circular tānoa ‘ava. Krämer mentions ifilele (Intsia bijuga - the Fijian vesi and Tongan fehi) and pau (Sapota achras) as being the most commonly used wood types (Krämer 1994, v. II: 244). Erskine (1853: 46) mentions the use of fetau (Calophyllum inophyllum - the Fijian dilo and Tongan feta’u), a sacred tree that was also used in Tonga, the Society and the Marquesas Islands for important objects such as bowls, canoes and headrests (Mu-Liepmann and Milledrogues 2008: 25). The villages Falealupo and Asau on Savai‘i were well known production centres for tanoa‘ava bowls (Mallon 2002: 17).
The majority of tānoa‘ava that entered predominantly German collections in the 1880ies have a diameter of 35-50cm, are metal tooled and invariably surrounded by a flat horizontal rim from which the interior abruptly falls away (Fig. 12). Their four legs are often just slightly tapered and lift the bottom of the bowl some 20cm off the ground giving it a somewhat suspended look when viewed from the side. At the end of the 19th century many-legged tanoa‘ava started being produced and quickly replace the older four-legged type. According to Te Rangi Hīroa, the additional legs were the result of a growing tourism in Sāmoa. Tourists were charged according to leg number, which increases with the size of the bowl (Buck 1930: 150). The many-legged Sāmoan bowls can have a distinctive small lip that extends the flat rim horizontally. The introduction of many legs left less space for the lug, which became a longer and narrower version of what has often been called a V-shaped lug. Rather than being rounded, the upper part of the legs, or even the entire legs, are sometimes squared. 20th century examples, stimulated by the tourist traffic, can have the flat rim area incised and filled with lime. Despite the production of many-legged bowls for sale to tourists, they were and still are used for actual ‘ava consumption by Sāmoans.
Figure 12: An older tānoa‘ava (Museum für Völkerkunde Dresden 48685) that was given by Chief Tamasese to the German consul Dr. Oskar Stübel in the 1880s. It shows the clear distinction between the flat rim and sloping inner walls of this comparatively shallow bowl. Right an example (Te Papa Tongareva Museum FE011948) collected in 1875. It typifies the many-legged broadly rimmed tānoa‘ava that became popular in the later 19th century. Its stained bowl indicates the bowl was in use before being turned into a painted and non-functional souvenir piece.
When it comes to lenticular ‘umete, neither lug nor leg shape allows us to clearly distinguish them from Fijian or Tongan examples. The legs are tapered and rather than being fully rounded they are sometimes keeled on the outside. They have a central ridge on their lower side running from tip to tip. Te Rangi Hīroa relates how in Savai‘i legless lenticular bowls with a flat bottom were used for ‘ava consumption (Buck 1930: 150).
Suspension lug shapes
Beside rim profile, the ever-present suspension lug of kava bowls can be another interesting feature to study. Lugs were essential for attaching the plaited sennit coir to hang the bowl on the wall and thus keeping its inside soot and dust free - the tapu character of bowls necessitating a circumspect and respectful treatment and storage.
In Fiji the lug is generally called mata (eye, face, front of something), in Lau daliga (ear) or sau, the latter also designating the white cowrie shells that can be attached to the sennit cord. Both mata and sau also refer to something that is perforated. In Tonga the lug is referred to more prosaically as taunga (hanger). The evolution of the suspension cord into an elaborately plaited sacred cord (wātabu or wā ni tānoa) embellished with white bulidina (Ovula ovum) shells, a symbol of godliness, is a Fijian innovation that was first documented at Bau in 1838 by Dumont d’Urville (Clunie 1986: 172).
The great number of provenanced yaqona bowls collected in Fiji allows a more thorough study than the fewer and mostly later examples collected in Sāmoa, not to speak of the only handful of Tongan ones. Similar to rim profiles Fijian mata types are a mixture of indigenous, as well as imported and transformed forms. They can be traced back to very simple square or trapezoid forms, sometimes notched in two or more places. They bear a strong resemblance to the salue (knobs) that ran down the middle of the fore and after deck covers of plank-built Sāmoan va‘aalo (bonito fishing canoes), where they were used to attach egg cowries (pule) (see Haddon and Hornell 1936: 236, Fig. 166). It is conceivable that in Fiji twin-notched mata of this type evolved into the dominant M-shaped form (Fig. 13 left column). On some later and large many-legged tānoa from Kabara the side bars are detached and have almost turned into legs. The side bars can also be absent, leaving just the middle part that has been described by Thompson as a V-shaped lug (Thompson 1940: 187). The term V-shaped lug, however, might more properly apply to a form that lacks vertical sides (Fig. 13, middle column).
Semi-circular lugs (Fig. 13, middle column, bottom) bear close resemblance to the perforated leads (sau, Tongan hau) through which the running stay of the Micronesian rigged Tongan/Fijian sailing canoes was passed (see Haddon and Hornell 1936: 308, Fig. 225). These particular vessels were built by the Lemaki in Lau as a replacement for the older sailing canoes such as the Polynesian-rigged tongiaki, which, lacking running stays, had no need of sau. This could date this particular shape to the late 18th century. Since it occurs only on very few bowls it seems that this lug shape was quickly replaced by the M-shaped type. A purely Fijian variant form of the M-shaped is illustrated in the right column of fig. 13. Rather than facing outward, these face downward clinging to the underside of the bowl, forming a decorative feature visible when the bowl is hung on the wall.
Figure 13: Left: A possible evolution of the M-shaped mata (frontal and top view). The last type is still produced today on Fijian tānoa bowls. Middle: V-shaped and semi-circular mata types. Right: The adhering M-shaped mata.
When comparing lugs of Sāmoan tānoa‘ava with their Fijian counterparts, one must remember that the majority was collected in the late 19th century, whereas some surviving Fijian yaqona bowls were most likely made in the late 17th or early 18th century. An early bowl collected by German resident Dr Bernhard Funk has a trapezoidal lug (Fig. 14 left) similar to Fijian mata. The association with Sāmoan va‘aalo bonito fishing canoes could make it a Sāmoan type that was subsequently transferred to Lau and Fiji. The absence of M-type lugs on Sāmoan bowls reinforces the suggestion that they are a purely Fijian, Lauan or Tongan evolution. Larger 19th century Sāmoan bowls with a flat rim are metal-carved and their lugs are more geometric and stylized (Fig. 14, left). Their sides are vertical and some have a cut-off tip of the chevron, a feature that is absent in the Fijian corpus. T-shaped Sāmoan lugs clearly relate to the more fluid T-shaped lugs of some older Fijian bowls (Fig. 14, right).
Figure 14: Left: Metal carved suspension lug types from four-legged and flat-rimmed Sāmoan bowls collected between 1880 and 1906. Right: T-shaped suspension lug types: The first lug is from a Fijian draunibaka, the second from a small tānoatavatava, both collected in 1875. The third is from a flat-rimmed Samoan tānoa‘ava collected before 1889.
The small number of provenanced Tongan kumete kava makes it impossible to get a representative sample of lug shapes comparable to Fijian and Sāmoan bowls. Many show both Fijian and/or Sāmoan influences, such as the M-type lug, chevroned fronts as well as trapezoidal or semi-circular shapes.
Figure 15: Lug shapes from Tongan kumete kava collected between 1774 and 1827.
Concluding remarks
In Western Polynesia bowls for the consumption of kava were anything but utilitarian vessels. They were valuables intricately linked to political, religious and economic systems. When carefully combing through 19th century sources it becomes clear that newly carved bowls were considered commodities that could be freely exchanged, whereas older bowls, which had accumulated their keepers’ histories and had enabled the communication with ancestor spirits and gods, were treasured items that could only be exchanged under exceptional circumstances. Many bowls still retain notches or semi-circular marks that testify to the many important occasions in which they were used and to the various generations of their keepers. The intermarriage of high-ranking Fijian, Tongan and Sāmoan lineages on the other hand allowed the geographic dispersal of these important heirlooms. The place of collection thus does not necessarily represent the place of manufacture. Only a differentiated and well sourced study and the comparison of distinctive features and technological subtleties allows us to trace their origins to some extent. On some tānoa collected in Fiji in the late 19th century one can observe the Samoan plank joining technique used to restore cracks or broken fragments, not only suggesting a Lauan origin, but also testifying to its use prior to entering the extended trade system sustained by Tongan seafarers.
The Lemaki craftsmen that had been moved to Lau by the Tui Tonga in the mid-18th century venerated the pulekula shell, a highly tapu heirloom orange cowry shell brought from Samoa. It was considered a tupua (ancestor/forbear) that embodied the Sāmoan goddess Lehalevao (Lyth, note 22 in Clunie 2013: 180). The stylized shape of that shell reappears in precious whale bone ivory beads of Tongan necklaces, the leads of drua and lugs of kava bowls, all produced by the same craftsmen. This underlines the importance and impact that craftsmen had in traditional oceanic societies. They were highly diversified, skilled, respected and sought-after members of the community whose work had a profound impact in all aspects of society, religion, exchange systems, the forging of alliances and the building of empires. More work needs to be done to search for these connections and to pay tribute to generations of craftsmen that written history has mostly forgotten.
References
Beaglehole, J.C., 1967. The Journals of Captain James Cook on his Voyages of Discovery. Vol. III. The Voyage of the Resolution and Discovery, 1776-1780. Cambridge: Hakluyt Society at the University Press.
Buck, P.H. (Te Rangi Hiroa), 1930. Samoan Material Culture. Bernice P. Bishop Museum Bulletin 75, Honolulu, Hawaii.
Clunie, F., 1986. Yalo I Viti (Shades of Viti). A Fiji Museum Catalogue. Suva, Fiji Museum.
Clunie, F., 1996. Une coupe sacrée à yaqona des îles Fidji – A sacred yaqona dish from Fiji. Art Tribal, Genève, pp. 3-18.
Clunie, F., 2013. Tapua: ‘polished ivory shrines’ of Tongan gods. The Journal of the Polynesian Society 122: 161-224.
Erskine, J. E., 1853. Journal of a Cruise Among the Islands of the Western Pacific: Including the Feejees and Others Inhabited by the Polynesian Negro Races, in Her Majesty's Ship Havannah. J. London, John Murray.
Gifford, E.W., 1929. Tongan Society. Bernice P. Bishop Museum Bulletin 61, Honolulu.
Haddon, A.C. and J. Hornell, 1975 [1936]. Canoes of Oceania, Vol. 1, The Canoes of Polynesia, Fiji and Micronesia. Bernice P. Bishop Museum Special Publication 27-29. Honolulu: Bishop Museum Press.
Herle, A. and L. Carreau, 2013. Chiefs and Governors; Art and Power in Fiji. Museum of Archaeology and Anthropology, University of Cambridge.
Krämer, A., 1994 [1902]. The Samoan Islands: An Outline of a Monograph with Particular Consideration of German Samoa. Volume I and II. Honolulu, University of Hawaii Press.
Labillardière, J.J., 1800. Atlas pour server à la relation du voyage à la recherché de La Pérouse. Paris, Jansen.
Lester, R.H., 1941. Kava drinking in Vitilevu, Fiji. Oceania XII (2): 97-121.
Mallon, S., 2002. Samoa Art and Artefacts—O Measina a Samoa. Honolulu: University of Hawai‘i Press.
Marshall, Y., A. Crosby, S. Matararaba, S. Wood, 2000. Sigatoka, The Shifting Sands of Fijian Prehistory. University of Southampton Department of Archaeology Monograph No. 1. Oxford: Oxbow Books.
Martin, J. (W. Mariner), 1827. An Account of the Natives of the Tongan Islands in the South Pacific Ocean. Volume 1 and 2, Third Edition, Edinburgh: Constable and Co. & Hurst, London: Chance and Co.
Mu-Liepmann, V. and L. Milledrogues, 2008. Sculpture: des œuvres anciennes aux créations contemporaines. Tahiti: Editions Au Vent des Iles.
Parham, J.W., 1972. Plants of the Fiji Islands. Suva: The Government Printer.
Thompson, L., 1940. Southern Lau: An Ethnography. Bernice P. Bishop Museum Bulletin 162, Honolulu, Hawaii.
Whistler A.W., 1991. The Ethnobotany of Tonga: The Plants, Their Tongan Names, and Their Uses. Honolulu: Bishop Museum Press.
Williams, T., 1982 [1858]. Fiji and the Fijians. Vol. 1: The Islands and their Inhabitants. New edition of original edition (London: Heylin) by the Fiji Museum. Suva: Oceania Printers Ltd.