Volume 102 1993 > Volume 102, No. 4 > Pulotu, Polynesian homeland, by Paul Geraghty, p 343-384
                                                                                             Previous | Next   

- 343

In this paper I shall attempt something so outdated as to be novel: to attribute historical truth to a myth. 1 Specifically, I shall demonstrate that it is possible to trace the mythical homeland of the Polynesians, Pulotu, to a location in Eastern Fiji. In so doing, I am continuing the investigations begun over 150 years ago by Horatio Hale (1846:119-20), the ethnologist/linguist of the American Exploring Expedition 1838-42, who argued convincingly that the Hawaiki of Eastern Polynesian mythology pointed to a real homeland in Savai'i, Samoa, a claim that has been generally accepted to this day. Hale also believed that the Pulotu of Western Polynesian mythology was, in fact, the homeland of the western, and hence the entire Polynesian people, but was unable to locate it with any certainty. 2


As Burrows (1938:73-6,120-2) demonstrated in his comprehensive survey of cultural traits, Hawaiki and Pulotu are in complementary distribution in Polynesia: Pulotu is “as definitely western Polynesian as the concept of Hawaiki as underworld or ancestral home is central-marginal [Eastern]” (Burrows 1938:76).

References to Pulotu in Tongan mythology can be found as far back as Anderson, who travelled with Cook on his third and final Pacific voyage, and noted in 1777 (Anderson 1967:949-50):

They have however very proper notions of the immateriality and immortality of the soul. They call it life, the living principle or what is more agreeable to their notions of it an O'tooa, divinity or invisible being, and say that immediately on death that of their chiefs separates from the body and goes to a place calls [sic] Boolootoo, the chief or god of which is Gooleho [Hikule'o]. They feign that this country was never seen by any person, is to the westward beyond Feejee (from whence one might suppose it is the most distant land they know, as many nations have plac'd their heaven in such a situation) and that there they live for ever, or to use their own expression are not subject to death again and feed upon all the products of their own country with which it is plentifully stor'd.

Mariner, in his classic first-hand account of early 19th-century Tongan society, adds many further particulars, including the following (Martin 1818:II:98,101): - 344

…mankind, according to a partial tradition, first came from Bolotoo, the chief residence of the gods, an island to the north-westward, and resided at the Tonga islands, by command of Tangaloa: they consisted of two brothers, with their wives and attendants, whose original they pretend to know nothing about … This island is supposed to be much larger than all their own islands put together, to be well stocked with all kinds of useful and ornamental plants ….

Gifford(1924:19,153-80,183)and Collocott(1921:231)recount a number of Tongan legends in which Pulotu is depicted as a distant land, in one case situated beyond Fiji, in another beyond Samoa, ruled over by the tyrannical Hikule'o, also known as Havea Hikule'o. Gifford (1924:19) describes Hikule'o in one tale as being blind, in two as a female. Collocott (1921:152) concurs that Hikule'o is usually considered female, but adds that Havea is the name of Hikule'o's mother. Tongan heroes and heroines travel to Pulotu and outdo the natives in all manner of competitions, or are sent by the King of Tonga to bring back the house of Hikule'o, or breadfruit (Artocarpus altilis) wood for the King of Tonga's house. In some cases they return with valuable products, mostly the kahokaho yam (a chiefly cultivar of Dioscorea alata), though others mentioned are taro (Collocasia esculenta), the kaumele (sic, for kaumeile) yam (a valued cultivar of Dioscorea alata), and a small rabbitfish ō (Siganus sp.). Two of Gifford's legends refer to Pulotu as Pulotu Tu'umau (firm or permanent Pulotu) without explanation (see note 12).

Futa Helu (personal communication) also notes that there are a number of old roads on different islands in Tonga called Hala-ki-Pulotu (Road to Pulotu). All lead in a westerly direction, and all end at points on the coast facing Fiji.

One of Gifford's tales differs from all others in claiming that “the path to Pulotu is between Eua and Kalau islands” (Gifford 1924:19), i.e., that Pulotu lies south-east of Tonga. Pritchard(1968 [1866]:397-8, 401) also claims an eastern location for Pulotu, while stating that spirits begin their journey there from the west end of Tongatapu. St. Johnston (1918:31), likewise, mentions that “several Tongans have told me of a rumoured ancient Bulotu at Eua island, and the Tongans would seem to have more likely come from the east than from the “negroid stock” to the north-west”. Whatever the significance of these accounts, it is plain that they represent a minority opinion (see Fig. 1).

Pulotu appears to have had a similar meaning in East Futuna (Burrows 1936:103; Andersen 1928:419) and East Uvea (Burrows 1937:85), but elsewhere in Polynesia was known only in Samoa. 3

In Samoa, Pulotu is also the term for an abode of departed spirits. The general entrance is situated at Le Fafā, a circular basin among the rocks at the west end of Savai'i, the most westerly island in Samoa (Turner 1861:235). Connected

- 345
Figure 1. Map showing Western Polynesia locations referred to in text.

with Le Fafā are two deep holes: “O le Lua-loto-o-Alii, or deep hole of chiefs, by which they passed to Pulotu; the other, O le Lua-loto-o-tau-fanua, or deep hole of the common people, by which they passed to Le nu 'u-o-nonoa, or the land of the bound, which is simply another term for the much-dreaded Sā-le-Fe'e” (Stair 1897:217). Pulotu is ruled by a cannibalistic ogre whose name, Saveasi'uleo (Turner 1861:237, Hovdhaugen 1987:190) is cognate with the Tongan Havea Hikule'o. Schultz (1980:57, 107, 110) describes Saveasi'uleo as half-man and half-eel. Turner (1861:259-260) gives the following description:

The upper part of his body was human, and reclined in a house in company with the chiefs who gathered around him; the lower was piscatorial, and stretched - 346 away into the sea. This royal house of assembly was supported by the erect bodies of chiefs who had been of high rank on earth, and who, before they died, anticipated with pride the high pre-eminence of being pillars in the temple of the king of Pulotu.

According to Hovdhaugen (1987:32-42), Saveasi'uleo ate all of his younger brothers, with the exception of the youngest, Ulufanuase'e. A daughter of Ulufanuase'e, Tilafaigā, eventually married her uncle Saveasi'uleo, and one of their daughters was Nāfanua, the famous war goddess. As in Tonga, there is a version, recorded by Stair (1895a:47), which implies that “Saveasi'uleo” is not one being but two, Saueā and Si'uleo, both children of Tagaloa-lagi.

Samoan legends are more specific than Tongan about Pulotu as a place of origin. The first to appear in print was in Turner 1884:222-3:

Of the [Samoan] group generally, it is said that a couple lived at Pulotu called Head of Day and Tail of Day. They had four children — (1) Ua, or Rain; (2) Farī, Long grass; (3) Langi, Heavens; (and 4) Tala, or Story. The four went to visit Papatea. Pulotu is in the west, Papatea in the east. The Papateans heard of the arrival of the four brothers and determined to kill them. First, Ua was struck on the neck; and hence the word taua, or beat the neck, as the word for war. This was the beginning of wars. Others stood on the neck of Farī, and hence the proverb in war: “Tomorrow we shall tread on the neck of Farī.” Others surrounded and spat on Langi, and hence the proverb for ill-usage, or rudely passing before chiefs: “It is spitting on Langi.” Tala was spared, and escaped uninjured.
Tala and Langi returned to Pulotu and told about their ill-usage. Then Elo, the king of Pulotu, was enraged, and prepared to go and fight the Papateans. This was the first war in history. They went, they fought, they conquered, and made a clean sweep of Papatea; and hence the proverb: “Like the rage of Elo.” Also for a village destroyed in battle they say: “Ua faa Papateaina” — made to be like Papatea.
All who fled to the bush were sought and killed, only those who fled to sea escaped. A man called Tutu and his wife Ila reached the island of Tutuila, and named it so by the union of their names. U and Polu reached Upolu, and hence the name of that island by uniting their names. Sa and Vaila reached Savaii, united their names also, and, for the sake of euphony, or, as they call euphony “lifting it easily,” made it Savaii instead of Savaila.
Elo and his warriors went back to Pulotu. Langi and Tala after a time came to Samoa, but went round by way of Papatea, and from them also the people of Manono and Apolima are said to have sprung.

Elsewhere, Turner (1884:227) provides further particulars: “Upolu was said to be the capital of Pulotu. In a time of war a number of people fled from Pulotu, - 347 reached this island of the Samoan group, and called it Upolu, in remembrance of their native land”. The essence of these stories is that Samoa was settled by fugitives from wars in Papatea and Pulotu. 4

The most recently published version of this story is in Hovdhaugen (1987:43-51) and is accompanied by a discussion of various versions which have been published over the years, most of which vary little from Turner's. A recent version not mentioned by Hovdhaugen, Tuiteleleapaga (1980:13-16) is also similar to Turner's, but states that Pulotu was not the name of Elo's island, but of a “suburb of the king's place or residence”. A significantly different version appears in Herman (1955:13), in which Pulotu is a group of islands east of Atafu and Elo is not the king of Pulotu, but a famous hero. After the war, the victorious inhabitants of Pulotu settled the Samoan islands and gave them their names. In the Samoan context, Atafu is understood to refer to the small island of that name in the Tokelau group, north of Samoa, so this account places Pulotu north-east of Samoa, in apparent contradiction of the general belief that Pulotu lies to the west of Samoa. I hope to demonstrate below that there is in fact no contradiction, because the Atafu of this legend is not Atafu in Tokelau.


The general direction indicated in the Polynesian descriptions of Pulotu suggests that, if it is or was a real place, it might be located somewhere in or around Fiji. Strengthening this proposal is the fact that the supreme deity of Eastern Fiji, Degei, is conceived as half-serpent (Waterhouse 1866:356), while the ruler of the Polynesian Pulotu is described in some accounts as being half-eel. 5

But when we search for a Fijian equivalent of Pulotu as a place of origin, we draw a blank: there has been some relatively recent prehistoric immigration, mostly from Tonga but, current popular beliefs notwithstanding, there is no Fijian tradition that their ancestors originated from anywhere outside Fiji (France 1969:4). 6 Similarly, for Pulotu as an abode of departed spirits, the most widespread functional equivalent is Bulu (Hale 1846:54, 18;, Waterhouse 1866:406-13; Williams 1870:204-9; Hocart 1929:184), which is conceived not as an island in the west, but as an underground or undersea world. 7

There is, however, a mythological place in Fiji named Burotu, which is clearly cognate with Polynesian Pulotu; but it is neither a place of origin nor an abode of departed spirits (Hale 1846:56; Waterhouse 1866:406-9). 8 Burotu is a paradise world full of beautiful things, proverbial for complete happiness: “I am in Burotu” means “I am thoroughly contented”. Although Burotu is the name of a number of localities and kin-groups in Fiji, 9 the Burotu of myth is generally agreed to be located in the area of Matuku, the southernmost of the islands of Yasayasa Muala 10 (the Muala group).

- 348

According to Fison (1907:16), “the Matuku people say that sometimes burnt-out fishing torches of a strange make, with handles of shell, drift ashore on their land, and when they pick them up they say, ‘see the torches from Burotu!’” Hocart (1929:195) mentions also that, “if the women of Matuku go out at low tide, at dawn they find red reeds that come from Mburotu”. Our more intimate knowledge of Burotu is due to the handful of valiant souls who have returned from visiting Burotu, in body or spirit, and told of it, though forbidden on pain of death to do so (St. Johnston 1918:42; Heatly 1923:11; Hocart 1929:211). Thompson (1940:116) reported that the people of Kabara say: “in olden days it was above the water but now it is submerged. The trees, like everything else on this island, are red so it is called Burotu Kula (Red Burotu). The island is inhabited by beautiful women so it is called vanua ni alewa (land of women). They are clean and light-skinned like Tongans and Samoans. The women are spirits who never die. Many men have tried in vain to find Burotu”. A specific instance is related by St. Johnston (1918:30): “The tale of the island Burotu ran north to Bau, and so inflamed that adventurous spirit Ratu Mara [half-brother to Cakobau, the most powerful Fijian chief of the mid-19th century] that, like a knight-errant of old, he started forth and swore that he would find Burotu or perish in the attempt. As a matter of fact he did neither, but the story at all events shows how earnestly he believed in it”.

All seven villages in Matuku claim descent, in whole or in part, from people of Burotu (Mocevakaca n.d.; Cagileta n.d.; Coriakula n.d.; Tikoisuva n.d.; Qio n.d.), though three are generally acknowledged to have a particularly close relationship — Makadru in the south-west, Levukaidaku in the south, and its offshoot Qalikarua in the east.

The people of Makadru are said to be descended from Loanimatalevu and his wife Di Vuna, both of whom came from Burotu (Tikoisuva n.d.). One of the clans of Makadru is named Burotu and it is said that until recently all their first-born offspring were female, as is the population of Burotu, and that Burotu appears whenever they teach meke (dances), though not only then. When Burotu emerges, it is seen as a large island beyond Daveta i Daku, a passage in the reef midway between Makadru and Levukaidaku, opposite Naitūtūraga, the site where Makadru was previously located, and appears to stretch westward towards Kadavu. It is believed that Burotu submerged because the chief of Makadru, Tui Daku, killed a white pig that was sent from Burotu.

The inhabitants of Levukaidaku say they smell the sweet-scented blossoms of Burotu when the wind is in the right direction, and frequently find coconut spathes and other debris of a deep red colour washed up on the beach. There is a stretch of sand near the village, named Nukusēlala, which emits the perfume of Burotu when disturbed, or at low tide in the morning, and at a sacred place

- 349
Figure 1. Map showing Fijian locations referred to in text.
- 350

called Vakavoleka, a stone causeway emerging from the sea is said to connect with Burotu (Tagilala 1989).

The inhabitants of Qalikarua, on the east coast of Matuku, which was settled from Levukaidaku (Cagileta n.d.), say that Burotu emerges somewhere between their village and Levukaidaku. One account, possibly European influenced, states that it submerged following an earthquake, the small offshore island of Nayanuyanu being a surviving remnant. The same earthquake is said to have resulted in the inundation of what is now the bay of neighbouring Totoya, originally a volcanic crater. On the island of Nayanuyanu there stands a magical tree called the duibana (different branches) which is said to be a composite of branches of different scented trees, and to have sprung from the garland of a female spirit, Yadi Vusonitokalau, who dropped it there in her haste to return to Burotukula before daybreak.

There are also numerous oral traditions connecting Kubunavanua, 11 ancestor of the chiefly families of Lau and, some say, Cakaudrove on Vanualevu, with Burotu. Sahlins (1962:234) tells of the Muala tradition that Kubunavanua went to Tonga and then came back to Burotu, accompanied by his three sons, and thence to Totoya. Thompson (1940:116) heard in Kabara that he “lived on the island [Burotu]. He had two canoes. He went to Totoya in one while his two sons went in the other to Moala. When Kubunavanua left Burotu the island disappeared”. The late Roko Sau (High Chief) of Totoya, Rusiate Sogotubu, told me that one of Kubunavanua's sons, who was born on Burotu, was named Rāvouvou ni [Prince of] Burotukula, and became known as Tuivanuakula (chief of the red or golden land) when he went to Nayau and founded what was to become the leading family of Lau. He added that Kubunavanua left Burotu for Totoya after returning from Tonga.

Matuku people are proud of their reputation as itaukei (owners) of Burotu. They explain alleged sightings of Burotu in other parts of Fiji (Heatly 1923:10)—notably near Ono in Southern Lau (Fison 1907:16), between Cikobia and the Yasawa islands (Barker 1925b:32-4), and south-east of Kadavu (St. Johnston 1918:30)—as being of Burotunawa (floating Burotu), and contrastingly refer to their own Burotu as Burotukula. 12 The suffix-kula emphasises the fact that everything there is bright red or intensely coloured, like the feathers of the kula lory (Phigys solitarius) that formerly served as a decorative border for mats, and were eagerly sought after by Tongans, for their own use, and as an item of commerce with Samoa (Kaeppler 1978:249, 252). The word kula is, of course, widely used in Polynesia for red feathers and other symbols of wealth and power, and shares many of the connotations of the English word “golden”. I shall return to the possible significance of this word below.

- 351

In addition to Matuku, many places in Fiji, though probably only eastern Fiji, have legends relating to Burotu. For example, Waterhouse (1866:406-8) tells of Tinanivatu, a lady of Nayaukumu on Ovalau, who married in Gau and was taken from there to Burotu. Her son was born there and returned to Nayaukumu as an adult, bringing with him as a present a small white cowry shell, which is still found on the reef there. None of these other legends claims that Burotu is close, or even suggests a location. The only island apart from Matuku which seems to have a close relationship to Burotu is Tuvana-i-rā, one of a pair of small islands situated south of Ono, the southernmost permanently inhabited island of the Lau group, approximately 350 km west of Ha'apai, Tonga, and 220 km south-east of Matuku. The Tuvana islands are usually either uninhabited, or populated by at most a family or two from Ono.

A well-known legend, first published in the official Fijian-language monthly Na Mata (Niu Madu 1894), tells of an Ono man named Tui Liku who was marooned on Tuvana-i-rā, location of the departure point of souls of the dead (icibaciba) of two of Ono's four villages, Matokana and Doi. There he met his clan's ancestor god, Ligadua, who invited him to accompany him, in spirit form, to Burotu. They left from Wailioci, and Tui Liku was told not to eat anything or enter any house if he wished to return to his body. This account ends rather abruptly with their arrival at a house in the village of Burotu. A more complete version of the same legend appears in two roughly contemporaneous accounts (Hocart 1929:210-1; St.Johnston 1918:31-46). Both agree that Tui Liku returned from Burotu with a number of useful plants which he planted at Tuvana-i-rā, St. Johnston specifying Matani wai as the place where they were planted. The plants included the dwarf coconut (niu leka), another variety or two of coconut (a coir coconut and king coconut according to Hocart, red coconut (niu rerega) according to St. Johnston), and the red ti plant (qai kula, a cultivar of Cordyline fruticosa). Hocart adds the red sovisovi plant (which I cannot identify), and St. Johnston includes the miji, “the little bird that flits among the coco-nut trees”. 13 St. Johnston gives the name of the departure point for Burotu as Vunikatavatu, rather than the Wailioci of Niu Madu's account, and adds a number of details that are not important for our purposes. I recorded the same story in 1979 from the lips of the late Tui [chief of] Matokana, one of the four villages of Ono, and his version (Sukanasau 1979) is almost identical with that published by St. Johnston; the useful plants Tui Liku brought back from Burotu he lists as the niu magimagi (coir coconut), niu rerega (red coconut), and qai kula (red ti plant, a cultivar of Cordyline fruticosa). The icibaciba (departure point of souls of the dead) of Tuvana-i-rā, from which Tui Liku left for Burotu, is situated on the north-west corner of the island.

- 352

Another well-known legend relating Burotu to Ono concerns the Levuka people of Lakeba, and was first recorded by Vakatovolea (1896), and subsequently by Fison (1907:12-7), Hocart (1929:208), and the Native Lands Commission (Tuinacevan.d.:3-4). The Levuka people, fishermen and potters who also taught meke (dances), captured and bound the lewa matagi, as the two sirens of Fijian mythology are known in Lau, while on a voyage to Ono. Arriving at Tuvana, they went ashore, leaving the captives in the charge of the children. By promising to teach them new dances, the lewa matagi persuaded the children to release them, then climbed to the masthead and sang this song:

Tuvana-i-colo Tuvana-i-rā Upper Tuvana, Lower Tuvana
Ko Nasali ka coka bābā Nasali stretches across
Ko Burotu e sa vuni tū gā Burotu is still hidden

They then stamped on the canoe, sank it, and fled, leaving the children to drown. This story certainly suggests a close connection between Tuvana and Burotu, but unfortunately offers no clue as to the location of Burotu. However, a version I recorded from Ono (Soko, personal communication, 1991) is a little more explicit, giving the third line as: Ko Burotu e sā vuni nō i rā, “Burotu lies hidden in the west”. Matuku is not due west from Tuvana, but its north-west direction can be readily accommodated by the Fijian rā. 14 So, while the Ono people claim proximity to Burotu, the direction indicated in these legends is still compatible with Matuku.

The most likely source for the claims of Ono with respect to Burotu is that alluded to above, in note 8. Whereas the belief that Matuku is the earthly counterpart of Burotu is widespread in Fiji, the tradition of Burotu being close to Ono and Tuvana, and of being the abode of departed spirits, is confined to Ono, and may therefore be presumed to reflect the partial continuance of a tradition inherited from the earlier Tongan occupants of Ono, and reinforced by the continued intercourse between Ono and Tonga.

Confusion between the rival claims of Matuku and Ono may have led Thompson (1940:116) to report that “Kabara people say that Burotu is an island situated between Ono and Matuku”. However, on the same page, she states that Moto, her main informant from Udu village on Kabara, “says he has heard that the souls of the land people went from Qaralevu to the highest rock on the reef, which is shaped like the roof of a house. He has not heard whether or not the souls of the land people go to Burotu, but he says that this rock faces Burotu and is connected with it by means of a submarine coral formation.” Qaralevu faces west, towards Matuku, whereas Ono is south of Kabara, so once again, Matuku is indicated.

- 353

It has often been asserted (Hale 1846:56, 195; Barker 1925a:25; Burrows 1938:121; Capell 1941:24) that Fijian Burotu is borrowed from Polynesian Pulotu. Linguistically this is improbable, as will be argued in the next section. Moreover, if the Fijians had borrowed the concept of Burotu from the Polynesians, it would be hard to explain why it is not considered the abode of the departed spirits, and why it is located specifically in Matuku, and not in some distant island to the north-west of Fiji.

Most of the evidence so far, therefore, from both Polynesia and Fiji, points to Pulotu/Burotu being located in or near Matuku. I now intend to strengthen the argument for such a location by proposing identifications for two lands that are associated with Pulotu in Samoan mythology — Papatea and Atafu.


Turning now to the islands associated with Pulotu in Samoan mythology, we note that Papatea is nowhere near so well known in Polynesia as is Pulotu, knowledge of it being confined to Samoa. The earliest and most detailed account is by Turner (1884:201-2), who states that the people of Papatea were required by the sun to furnish daily human sacrifices, and that a brother and sister fled from there to Manu'a. Turner's account (1884:222-3) of the war in which Papatea was defeated by Pulotu has already been quoted in full, but it will be useful to recall the references therein to Papatea:

The four [brothers from Pulotu] went to visit Papatea. Pulotu is in the west, Papatea in the east. The Papateans…determined to kill them. [In revenge, the Pulotu army] went, they fought, and made a clean sweep of Papatea… All who fled to the bush were sought and killed, only those who fled to sea escaped [to Samoa]… Langi and Tala [from Pulotu] after a time came to Samoa, but went round by way of Papatea.…

Krämer's discussion (1902:I:23, 401) of Papatea covers much the same ground, repeating that Papatea is a spirit abode lying far to the east. Hovdhaugen (1987:187) similarly defines Papatea as “a mythical island or group of islands situated in the East”. Herman's version (1955:13) is essentially the same as Turner's.

As for the actual location of Papatea, there are a couple of candidates in the literature. According to Smith (1910:114), “It is said that Papatea is the Samoan name for the Marquesas Group, but further evidence of this is wanting”. Further evidence is still wanting, but it is possible that Smith's informant confused the Marquesas with the Tuamotus: as Turner (1884:223) speculated, it appears that Ma'atea in the Tuamotus may have been previously known as Papatea (Beagle- - 354 hole 1968:293). 15 However, if Papatea is indeed an old name for Ma'atea, it would seem to predate 1770, since not only Parkinson in 1770 but also the Spaniards who occupied Tahiti in the early 1770s recorded the island's name as Matea (Dening 1962:136). It is also given as Ma'atea in an ancient Tahitian chant recorded by J. M. Orsmond in the early 19th century (Smith 1898:2-3) and in the presumably ancient “log-book” of Tuamotu migrations which Smith obtained in 1897 (Smith 1910:120). In any case, given the widespread Polynesian words papa “flat hard surface” and tea “white”, it is hardly surprising that this name should be applied to an island of raised coral. 16

Another candidate is found in the celebrated chart of islands known to Tupaia, the renowned Tahitian navigator who sailed with Captain Cook in 1770, which shows an island named o-Papatea. Smith (1898:13) remarks as follows: “Appears from the chart to be somewhere in the neighbourhood of Suwarrow Island. It is mentioned in the Samoan traditions”. Dening (1962:135) does not attempt an identification of o-Papatea, but identifies two of its closest neighbours on the chart, Whennua-oora and o-Poppoa, as Fenua Ura and, probably, Pukapuka, which would place o-Papatea in the middle of the open ocean, but nevertheless close to Suwarrow and east of Samoa. Therefore, although the exact location is in doubt, both of these judgments based on Tupaia's chart accord well with statements from Samoan mythology that Papatea is situated to the east of Samoa.

But this is far from being the only interpretation of the location of o-Papatea according to Tupaia's chart. As Hale (1846:122-3) has pointed out, all islands unknown to Cook and his officers are reversed on the north-south axis. In the case of Papatea this consideration is of little importance, since it is shown on the chart lying just a little north of due west of Tahiti. Smith (1898:11-2) has also shown that two or three islands are reversed on the east-west axis, but none of these lies in the vicinity of o-Papatea. While the chart, allowing for these discrepancies, seems to be fairly accurate as to the direction indicated from Tahiti, with respect to distances it appears to be considerably less so. It is true that some of the islands lying in the same general direction appear to be located relative to each other, but there is no overall scale. For instance, Raro-toa (Rarotonga) is located at twice the distance from Tahiti as Oheavai (Savai'i), when in fact Savai'i is about twice as far away as Rarotonga. So the location of any unknown island is open to a number of different interpretations, depending on which nearby known island is used as a reference. It could just as correctly be inferred from Tupaia's chart that Papatea is located ESE of Rotuma and SSW of Savai'i — in other words, in the Fiji group. Even if Papatea was located with reference to Fenua Ura, which is its closest neighbour on the chart, it may have been so placed not because of Tupaia's geographical knowledge, but because of his mythological knowledge of the proximity of Papatea to Pulotu, the Eastern - 355 Polynesian name for which is, as will be argued below, Fenuakura.

How, then, to explain the persistent claim in Samoan mythology that Papatea is a distant island to the east of Samoa? I believe that the persistence is only apparent, in that Krämer, Herman, and Hovdhaugen were simply following, and misinterpreting, the location given originally by Turner: “Pulotu is in the west, Papatea in the east”. This juxtaposition of opposites is a word-for-word translation of a Samoan construction equivalent to a comparative sentence (Hovdhaugen 1985:96), a type of costruction that is also found in related languages, e.g., Tuvalu (Besnier 1981:266). The Samoan sentence “America is large, Australia is small” means not that either is large or small in any absolute sense, but that America is larger than Australia. Similarly, the sentence “Pulotu is in the west, Papatea in the east” is a statement of their positions not relative to some absolute point, such as Samoa, but relative to each other. It does not mean that Pulotu is west of Samoa and Papatea east of Samoa, but that Pulotu is west of Papatea, and Papatea east of Pulotu. Locating Papatea east of Pulotu also removes a problem with the previous analysis: since Pulotu is agreed by all to be west of Samoa, if Papatea were east of Samoa, then its people must have made their long journeys to Pulotu with their eyes closed, in order to avoid discovering Samoa on the way.

It was noted above that Levukaidaku is one of the villages in Matuku closely associated with Burotu. The honorific name for Levukaidaku is Babasea, which, I suggest, is cognate with Papatea. 17


Atafu is one of the islands of the Tokelau group north of Samoa, but its location in mythology is unclear. Stair (1895a:48) reported that, “in what I imagine to be one of the oldest traditions I have obtained, “Atafu” is mentioned as the island or land from which one of the first parties of immigrants came”. In Krämer's (1902:I:8,403-5) version of the sun sacrifice story, the refugees from Papatea fled to Atafu and then to Manu'a. Herman (1955:18) states that Atafu is east of Manu'a, which hardly supports the Tokelau identification. He adds (1955:92) that people from Atafu came to quarry stone axes on the mountain behind Leone on Tutuila, but this is not particularly useful information, since adzes from the Tataga Matau quarry at Leone have been found all over the western Pacific (Best, Sheppard, Green and Parker 1992). Schultz (1980:124) also tells of the people of Atafu being oppressed by their king, the sun, and forced to make human sacrifices to him, but believes Atafu to refer to the island in Tokelau, since there is a tradition of people fleeing Atafu in Tokelau and settling in Malie on 'Upolu (Stair 1895a:48; Newell 1895:239). Both Newell and Schultz, however, give the name of the land these people fled as Atafu-mea. This raises the possibility that there are two places named Atafu. Since it is Atafu in - 356 the Tokelau group that is distinguished by the name Atafu-mea, and the suffix-mea meaning “brown or red” (Milner 1966) is indeed appropriate for an atoll, simple Atafu most probably refers to another island of that name which is not an atoll.

It will be recalled that, in the version of the Pulotu myth recorded by Herman (1955:13), Pulotu is said to be situated east of Atafu. Some 120 km west of Matuku lies the island of Kadavu, the fourth largest island in Fiji. This, I propose, is the Atafu homeland mentioned by Stair. 18 Unlike Pulotu and Papatea, Atafu and Kadavu derive from an etymon which is reconstructible as far back as Proto Oceanic, as POC *kadavu “raincloud”, though no such meaning is found in any Central Pacific language, i.e., in Fijian, Polynesian, or Rotuman. 19


Those not aware of recent developments in Proto Central Pacific phonology may be reluctant to accept the three comparisons proposed above — Fijian Burotu, Babasea, and Kadavu with Samoan Pulotu, Papatea, and Atafu respectively — since each pair contains a sound correspondence that would have been considered, until fairly recently, problematic. However, evidence presented in Geraghty (1986:295-6,301,305), and elaborated below, allows us to reconstruct *burotu, *babajea, and *xadavu at the PCP (Proto Central Pacific) level, i.e., as words used by the ancestors of the contemporary speakers of the Fijian, Polynesian, and Rotuman languages at the time when they still spoke essentially the same language. In the data below, forms reconstructed for PCP are marked with a preceding asterisk. 20

Reconstructing *burotu

Fijian/r/usually corresponds with PPN (Proto Polynesian) *r, but frequently with PPN *l, as in the following examples:


  • *riri boil: PFJ *riri, PPN *lili
  • *rau-leaf: PFJ *rau-, PPN *lau
  • *rogo quiet, silent: PFJ *rorogo, PPN *logo


  • *marari k [kind of] fish, wrasse: PPN *malali (PMC *merari)
  • *gara scream, howl: PFJ *gara, PPN *gala
  • *turu drip: PFJ *turu, PPN *tulu
- 357

It is unlikely linguistically that Fijian Burotu is borrowed from Polynesian Pulotu, since Tongan /p/ is normally borrowed into Fijian as /p/ or /v/, e.g.:

  • TON kapa sheet metal — FIJ kapa, kava
  • TON pōtea k scented coconut oil — FIJ vōtea
  • TON puaka pig — FIJ puaka, vuaka
  • TON pulelulu Wednesday — FIJ vulelulu (obs), vukelulu
  • TON tupenu wrap-around cloth — FIJ tupenu, tuvenu

and Tongan /l/ is normally borrowed into Fijian as /l/, e.g.:

  • TON puleluluWednesday — FIJ vulelulu (obs), vukelulu
  • TON talanoa tell a story — FIJ talanoa
  • TON telefua naked — FIJ telefua, televua 21
Reconstructing *babajea

PCP *j is the reflex of PEO *j as proposed by Geraghty (1983:149-53; 1986:298, 301) on the basis of the correspondence: Fiji /s/, PPN *t or *s, Rotuman *j, PSS (Proto Southeast Solomons) *d. 22 It approximates to the POC *nj proposed by Milke (1968), and the PCP *c of Blust (1976). Only in Rotuman is it retained as a distinct phoneme. In PSS it merges with the reflexes of *d and *dr, in PFJ with *s, and in PPN with *s or *t and *d. Given this pattern of mergers, it seems likely that *j was the “prenasalised” counterpart of *z (Geraghty 1986:297-300). PCP *j was probably phonetically [t∫] or [ts], like its only unique reflex, Rotuman /j/ (Churchward 1940:13,83). The evidence for PCP*j is presented below. A number of items included in Geraghty 1983 and 1990 only on the strength of external evidence, usually PSS *d, are omitted here.

As with PCP *r, *c, and *z (Geraghty 1986:295), the PPN split reflex is problematic. There is no obvious conditioning other than perhaps a slight tendency to *s before front vowels.


  • *jamu scraps of food: PFJ *sabusabu, PPN *samu, ROT jamjamu
  • *jamu(?)k(palm) fruit stem: PFJ *sāmoa, PPN *taume spathe (met.), ROT jamu?a
  • *jao spear: PFJ *, PPN *tao, ROT jao
  • *jaojao k fish, small Sphyraena: PFJ *sāsā, PPN *(st)ao(st)ao
  • *jara slip, slide: PFJ *sara, PPN *tala put on (clothes), ROT jara
  • *jau strike, beat: PFJ *sau strike, beat, tattoo, PPN *tatau tattoo, ROT jau
  • *je(?)k bird, Lalage maculosa, Polynesian triller: PFJ *sea, PPN *(hs)e(?)a, ROT jea
- 358
  • *je(?)(ei) k insect: PPN *se?e grasshopper, locust, ROT jei cricket
  • *jei tear, rip: PFJ *sei, (PPN *sae), ROT jei
  • *jeu k fish, Carangidae sp: PFJ *seu Gnathanodon, PPN *teu Caranx sp.
  • *jevu splash water: PFJ *sevu, ROT jehu drizzle
  • *jexejexe k fish, Arothron sp: PFJ *sekeseke, PPN *te?ete?e
  • *je?evo(ou) k fish, juvenile mullet: PFJ *sēvou, PPN *te?efō (TON)
  • * k plant, Cordyline: PPN *, ROT
  • *jiko- kingfisher: PFJ *sikorere Artamus, wood swallow, PPN *tikotara
  • *jila look sideways, squint: PPN *sila, ROT jila
  • *jila (canoe) sheet: PFJ *sila, PPN *tila yardarm
  • *jiji slip: PFJ *sisi, ROT jiji creep, crawl
  • *jobu div:, PFJ *so(bv)u descend, PPN *sopu, ROT jopu
  • *joli pick, gather: PPN *toli, ROT joli
  • *jona yaws, (octopus) sucker: PFJ *sona, PPN *tona, ROT jona
  • *jopu nod: PFJ *sopu, ROT jopu
  • *jou (sea) rough: PFJ *sou splash, PPN *sou, ROT jou ripple
  • *jove k shellfish: PFJ *sove barnacle, PPN *tofe k bivalve mollusc


  • *baja close together: PFJ basa meet, opposite, ROT paja
  • *duji point: PFJ *du(cs)i, PPN *tus(iu), ROT juju
  • *guju- mouth: PFJ *gusu-, PPN *gutu, ROT nuju (*g>n irregular)
  • *gujula k fish, k long-snouted Lethrinus sp: PFJ *gu(st)ula Lethrinus elongatus, PPN *gutula k fish, Lethrinus sp.
  • *gwajala k fish, Epinephelus sp: PFJ *kasala (for expected *gwasala), PPN *gatala, ROT vajala
  • *(jsØ)ije Hemirhamphidae, garfish: PFJ *sise, PPN *ise, ROT jija (*e>a regular) 23
  • *kaja- (kava) stem: PFJ *kasa, PPN *kata
  • *kiajo, *ikajo outrigger boom: PFJ *ikaso, PPN *kiato
  • *keju- back of head: PFJ *kesu-, ROT ?eju
  • *laja tame: PFJ *lasa, PPN *lata
  • *laje coral: PFJ *lase PPN *lase, ROT laje
  • *majaga- (road etc) fork: PFJ basaga- 24 PPN *masaga twin, ROTmajaga
  • *maja(?)u clever, expert: PPN mata(?)u, ROT majau
  • *miji suck: PFJ *misi PPN *miti + lick
  • *muju cut off: PFJ *musu break, PPN *mutu
  • *(?)oje k fish, Mulloidichthys sp., goatfish: PFJ *ose, PPN *(?)ose (MAE), ROT oje M.vanicolensis
  • *taji shave: PFJ *tasi, PPN *tasi, ROT jaji
  • *uja transport, carry (cargo): PFJ *usa, PPN uta
- 359
  • *xujim(ai), xumij(ai) crave fish or seafood: PFJ *kusima, PPN *?umiti
Reconstructing *xadavu

PCP *x was first reconstructed in Geraghty 1986:305. Its Fijian reflex is /k/, the same as that of PCP *k. As reported in Geraghty 1983:160-1, it is distinguished from *k by the reflexes PPN *? (or occasionally a *?/k doublet) rather than *k, and Rotuman /Ø/ or /?/ rather than /?/. It may be the result of an incomplete change, rather than an actual PCP phoneme, but its inclusion lends symmetry to the system, since the velar series now parallels the labial in containing a fricative as well as a voiceless stop and a prenasalised stop. The evidence for PCP *x is presented below. The scant Rotuman evidence offers little rationale for the split reflexes, /Ø/ and /?/.


  • *xa(bp)a (house) wall: PFJ *ka(bp)a, PPN *(?)apa(?)apa
  • *xana-N past: PFJ *kana-, PPN *?ana
  • *xanusi spit: PFJ *kānusi, 25 PPN *?anusi, ROT anusi (PN loan?)
  • *xarawa k large shark: PFJ *karawa Isurus, PPN *?alawa (EUV Carcharhinus limbatus; TVL C. longimanus; REN, TIK C. melanopterus; TVL, EP Negaprion)
  • *xata make mark, show up clearly: PFJ *kata, PPN *?ata shadow, reflection; (tattoo) bright (REN), ROT afa
  • *xavelu, vaxelu wipe anus: PFJ *kāvelu,25 PPN *fa?elu
  • *xoda eat raw (flesh): PFJ *koda, PPN *?ota
  • *xola split (firewood): PFJ *kola, PPN *?ola wedge, ROT ?olo chop, cut (a>o irregular)
  • *xota dregs, refuse: PFJ *kota, PPN *?ota, ROT mofa 26
  • *xōtai fruit salad: PFJ *kōtai, PPN *?ōtai
  • *xū-t bite off: PFJ *kū-t, PPN *?ū-t
  • *xua don't, cease: PFJ *kua, PPN *(k?)aua (/a/ irregular), ROT ?u?ua
  • *xujim(ai), *xumij(ai) crave fish or seafood: PFJ *kusima, PPN *?umiti
  • *xuru rumble: PFJ *kuru, PPN * ?u?ulu, ROT ?ū bang (loss of *r is regular)


  • *axa- jawbone: PFJ *yaka- mouth, PPN *a?a
  • *boxoi k pudding: PFJ *bokoi, PPN *po?oi, ROT po?oi (PN loan?)
  • *gwaxa k fish, Hemiramphus: PFJ *g(w)aka, PPN *ga?a Rastrelliger; PTK *mwaki juvenile Strongylura
  • *jexejexe k fish, Arothron, puffer: PFJ *sekeseke, PPN *te?ete?e
  • *matuxurau k fish, Parupeneus sp., goatfish: PFJ *matukurau P. porphyreus, PPN matu?ulau
  • *maxavu Magellan's clouds: PFJ *makavu, PPN ma?afu
- 360
  • *maxota k tree, Dysoxylum: PFJ *mākota,25 PPN *ma?ota
  • *noxa tie up, tether: PFJ *noka, PPN *no?a
  • *saxalo scrape (coconut): PFJ *i/sakalo coconut scraper, PPN *sa?alo, sākalo
  • *tānoxa, tāxona k wooden bowl: PFJ *tākona, PPN *tāno?a
  • *vaxa-V often: PFJ *vaka, PPN *fa?a
  • *vaxo peg: PFJ *vako, PPN *fa?o

The perfect phonological correspondence of these three place-names and the parallel relative locations of their referents in Fiji and in Samoan mythology leave little doubt but that they are cognate, real places that, through lack of contact, have lost their reality to the Samoans and become mythical. It is quite plausible therefore that the Samoan myths and beliefs we have been discussing are based on a historical incident in which refugees, and victors, from a war in or near Matuku in Fiji fled to Samoa, and settled there.

If this conclusion is correct, then, according to the partial myth recorded by Mariner, it would appear that the same people also fled to Tonga and settled there. The fact that Mariner describes Pulotu as much larger than all of Tonga need not necessarily be viewed as contradictory, since the people who fled to Tonga would have handed down memories of a homeland that comprised not only Matuku itself, but probably the whole of the Fiji group, including the comparatively vast island of Vitilevu.


While Kadavu and Babasea are real places, the precise location of Burotu remains a mystery. Its former inhabitants are known to live in at least parts of Samoa, Tonga, Matuku in Fiji, and in the various parts of Lau and Cakaudrove where the descendants of Kubunavanua are said to have provided the chiefly caste. Its location, according to Samoan sources, is east of Kadavu and west of Babasea, which is most probably south-east Matuku. Some sources from Matuku agree that Burotu lies between Matuku and Kadavu, others say it lies south or east of Matuku, while those from Kabara and Ono are less specific, but do not preclude the Matuku area as a general location. All Fijian sources state that Burotu is now submerged, and periodically re-emerges, but most accounts do not indicate that it submerged because of a volcanic eruption or similar catastrophe. We have noted above the Makadru belief that Burotu was submerged as a punishment for their killing the white pig sent from there. Other traditions state that it sank when the ancestor hero Kubunavanua stepped on it when leaving it for the last time, an explanation apparently supported by Thompson's (1940:115) statement that Burotu used to be known to the people of Ono as Butudromu (butu “step on”, dromu “submerged”). Mokunitulevu Na - 361 Rai (n.d.: 12) asserts that it disappeared gradually, beginning in 1760, when the Levuka people moved from Bau to Lakeba.

The explanation I wish to propose is that Burotu was a part of Matuku, most likely the western part as opposed to the eastern, or possibly the northern as opposed to the southern, or perhaps a nearby island. This accords well with the Matuku people's claims that they all “came” from Burotu. It may have been a political or commercial centre, possibly an important source of red feathers (hence the -kula suffix) which, like Camelot, acquired fantastic traits as it declined. In view of the Makadru belief that Burotu is apt to emerge when they teach dances, and the meanings of some of the Polynesian reflexes of Burotu (see note 3), it may also have been a centre for music, dance, medicine, and learning; and the current belief that Burotu women are fair may be the result of Tongan women marrying into Burotu, either as articles of exchange or in strategic marriages. The name Burotu is no longer applied to any real part of Matuku, other than the kin group in Makadru. The name Matuku is said to have been brought by a group of immigrants from the Waimarō area of Tailevu on Vitilevu, from whom the present Rātū (chief) of Levukaidaku is descended. There is indeed a mountain there named Matuku, near the present-day village of Naivicula.

The suggestion that Burotu may have been part of Matuku is supported by the etymology of what would have been its counterpart, Babajea. Baba-, probably related to baba “side, slope”, is a prefix used in Kadavu and Matuku for halves of an island or social unit. In Kadavu, Babaceva, the side of the island facing the ceva (south wind), contrasts with Babatokalau, the side facing the tokalau “north wind”, and the Matuku village of Levukaidaku is also divided into two social units named Babaceva and Babatokalau. Matuku as a whole is, likewise, divided into two parts: Babamatuku, comprising the villages whose chiefs are held to descend from the immigrants from the Waimarō area of Tailevu on Vitilevu, and Yavusavakavanua, approximately “the land tribe”. Babajea, therefore, may have referred to one side of the island of Matuku, in contrast with Burotu.

It is tempting to speculate that, if Burotu was indeed an important centre of the Polynesian red-feather trade, one of the reasons for the decline may have been that the kula was hunted to extinction. It is inexplicably absent from Totoya (Clunie 1984:58), and from certain other islands in Lau, such as Cicia and Vanuabalavu, though present in Matuku (Moseley 1944 [1879]:255; Beckon 1989). It is also tempting to speculate further that Matuku may have been an early source of feathers of Prosopeia splendens (Rinke 1989), the Kadavu musk-parrot, which is the only Fiji parrot with bright red feathers, the others being maroon or yellow, and bears the appropriate name kā-kula(being generic for “parrot”). Fiji parrot feathers were also sought after by Samoans and Tongans, in addition to those of the kula lory (Clunie 1986:150). Although the Kadavu musk-parrot is not found on Matuku today, an unspecified musk-parrot was present as far east as Lakeba - 362 up until at least 2250 B.P. (Best 1984:530). It is possible that the Lakeba parrot became extinct through a combination of over-exploitation for the red-feather trade, and loss of habitat after extensive burning of inland forests for farming that occurred about 2000 B.P. (Best 1984:563), resulting in a long period during which Lakeba was sparsely populated (Best 1984:643). Subsequently, Matuku and the other islands of Yasayasa Muala were resorted to, until the parrot again became extinct. 27

With the demise of Burotu as a commercial and cultural centre, the reality faded and the legend grew. The legend of a remote island paradise and source of exotic plants and animals which attached to Burotu appears to be ancient. Remote islands peopled only by women are reported from Pukapuka (Beaglehole 1938:401-2), New Britain (Dixon 1916:140-1), and the Trobriands (Malinowski 1929:356-9), where it is related that some men who succeeded in returning alive brought with them a new variety of banana. For the people of Vanikoro, Veluko is a non-geographic island between Vanikoro and Tikopia from which originated the Tamate spirits (Tua 1979:10). Panoi, the Hades of the Banks Islands, has trees with red leaves (Codrington 1891:275). In Micronesia, Goodenough (1986:554, 563) reports that a number of “ghost” islands occur in the sailing directions of Carolinian navigators, and that a Pohnpeian hero travelled to the mythical island Kachaw and returned with a kind of mangrove used in housebuilding. In Fiji, Barker (1925a:25-9) relates a legend from the north coast of Vitilevu of a visit to an island named Vanua Alewa (Land of Women) situated somewhere in the ocean north of Vitilevu and identical to Burotu in all but name.

Whether Burotu truly submerged is an open question. The story may be simply part of the inherited tradition of an island paradise; or it may be that Burotu was a coastal part of Matuku or a nearby island that was, in fact, inundated. Three plausible candidates are Lomaji Bay, a deep indentation on the western coast of Matuku; part of Nayanuyanu island, off the eastern coast of Matuku, which is mentioned in the Qalikarua tradition noted above; and the bay of Totoya, originally a volcanic crater, which may have been breached and inundated quite recently (Nunn personal communication). Sea-levels in Fiji were much the same as today about 4,400 years ago (Gibbons and Clunie 1986:59), then rose to between 1-2 metres higher about 2,000 years ago (Nunn 1990:297). During the period 4,400-2,000 B.P., therefore, land up to two metres above present sea-level could have been inundated; but, since I shall argue below that the real or metaphorical disappearance of Burotu is a more recent event, sea-level change was probably not a factor. However, it would be rash to rule out natural causes altogether. Nunn (1987:10) points out that the Yasayasa Muala area, which comprises Muala, Totoya, and Matuku, has “a long and continuing history of tectonic disturbance, which has even caused some of the coastal plains - 363 exposed by the slight fall of sea-level in the last 3,000 years or so to have been partially submerged”. There is also the ever-present possibility of inundation by tsunami.


There are indications that Burotu may have been known by another name: Vanuakula. The Fijian historian Mokunitulevu Na Rai (n.d.:12) states that both are names of the same place, and the name Vanuakula is appropriate as a description, being composed of vanua “land” and kula “red, golden”, given that Burotu is said by Fijians to be full of such coloured things. Furthermore, the son of the culture hero Kubunavanua, who was born in Burotu, became known as Tuivanuakula (tui “chief, monarch”). Other traditions state that Kubunavanua himself was also known as Tuivanuakula (Reid 1990:9, 20).

From the Polynesian side, Smith (1910:112) reproduces a list of lands of origin in the form of a recitation found in manuscripts which he brought from Rarotonga in 1897, thus: Atia-te-varinga-nui, Avaiki-te-varinga, Iti-nui, Papua, Enua-kura, Avaiki, Kuporu, Manuka. The last three are readily identifiable as Savai'i, 'Upolu, and Manu'a in Samoa. The Rarotongan Enua-kura is perfectly cognate with Fijian Vanuakula. Smith believes that Iti-nui refers to Vitilevu, the main island of Fiji, and indeed Rarotonga 'iti, often spelt without the glottal stop, is cognate with Viti, the Fijian name for Fiji, while nui translates Fijian levu “large”. So, whatever the interpretation of Papua, 28 Enua-kura appears to be the name of a place between Vitilevu and Samoa, i.e., somewhere in the islands of Eastern Fiji. A Rarotongan myth (Gill 1876:109-14) set in Samoa also tells of a visit by Tane to Enua-kura, and a Rarotongan genealogy (Anon. 1892:25) mentions Enua-kura as one of the places conquered by Tu-tarangi, the others that can be identified being all situated in the Fiji-Tonga-Samoa area. In New Zealand, Whenua Kura is said to be the name of the beach in Hawaiki from which the Te Arawa and Tainui canoes set off on their voyage to New Zealand (Stair 1897:242). From Takume, Tuamotu, Stimson (1932:181-90) records a chant recounting a voyage to a distant land named Henua-kura, either identical to or beyond Havaiki. Recall also that the problematic Papatea of Tupaia's chart is situated close to Fenua Ura. As noted above, it may be that Tupaia was aware of a historical connection between Vanuakula (i.e., Burotu) and Babajea, and so located Papatea near the only Vanuakula he knew, i.e., Fenua Ura of the Society Islands. In Hawaii, the cognate honua'ula is not the name of a distant place, but of a dark brown-red variety of sugar-cane, its name perhaps telling its supposed origin.

In Kiribati, Grimble (1972:181-2, 1989:229) tells of a “navigating clan” named Benuakura, clearly a loan from the Polynesian Fenuakura. Its totem is a red-feathered bird, reputed to be man-eating. Its founder was born in the - 364 mythical land of Roro, but lived in Samoa before coming to Kiribati. This, then, is further evidence, probably from Samoa, of a place named Fenuakura and associated with red-feathered birds. 29

St. Johnston (1921:97) also believed that Enua-kura refers to a place in Fiji, but specified Taveuni, since it “was said to be the island where the much-prized scarlet parrot feathers were obtained. Taveuni has been famous for these from time immemorial”. Nineteenth-century writers do indicate that Taveuni was then a major source of red feathers (Clunie 1986:150-1), but available evidence does not permit us to state that this had been so since time immemorial, even though it is clear from archaeological evidence that Taveuni has long been an important political centre. As argued above, the presence of the Kadavu parrot and relative proximity to Polynesia would have made the Matuku area an attractive proposition to Polynesian kula-hunters. 30


How do the findings of this paper fit in with other evidence for the settlement of Polynesia? The current consensus appears to be that archaeology (Best 1984:626-7, 653-4) and linguistics (Geraghty 1983:381) support the view that Polynesia was settled from somewhere in eastern Fiji, which is consistent with the Pulotu legends. The notion that Tonga was peopled before Samoa has been discredited archaeologically, the current view being that all Western Polynesia was settled at about the same time (Kirch 1988:244); 31 and, indeed, in both Tonga and Samoa direct origin is claimed from Pulotu or thereabouts. On the other hand, a number of considerations argue against associating the traditions respecting Pulotu, Papatea, and Atafu with the archaeologically determined initial settlement of Western Polynesia.

A major consideration is the timespan involved: there must be some doubt that an oral tradition can survive for some 3,000 years. There are a number of plausible migration traditions in Eastern Polynesia, for instance, that of the secondary settlement of Rarotonga from Samoa (Williams 1842:51-2), but the timespan involved is considerably less than 3,000 years. The earliest tradition known from Rotuma, that of the arrival of Raho from Samoa (Pritchard 1968 [1866]:379; Eason n.d.[1951]:2), is clearly not of the original inhabitants but of later arrivals from Polynesia. Similarly, the commonly held belief among the Kiribati people that Kiribati was settled from Samoa appears to refer to a relatively recent event (Grimble 1989:255). So, while the Pulotu story may indeed be the oldest settlement tradition extant in Western Polynesia, it is far from certain that it is as old as 3,000 years, and therefore unlikely that it tells of the initial settlement of Western Polynesia. 32

Another point of significance is the etymology of a word that has featured prominently in this paper: kula “collared lory, Phigys solitarius”, the main - 365 source of the red feathers so highly prized by Samoans and Tongans (Watling 1982:89). Although the name is widespread in Fiji, the linguistic evidence points to it being a loan from a Polynesian language. The ultimate source of this Polynesian word appears to have been Proto Oceanic *kurat “k tree, Morinda citrifolia”. 33 This well-known Pacific tree is famous for the red dye produced from its roots (e.g., Ivens 1927:195; Firth 1985:298; Pétard 1986:280-2), and its name throughout Fiji is kura, which is the regular reflex of *kurat. 34 Given the fact noted above that *r sometimes develops into PPN *l, and the obvious semantic connection between red dye and the various meanings of PPN *kula, 35 it could well be that PPN *kula reflects POC *kurat, and that the occurrence of kula in Fiji is the result of borrowing, the probable source being PPN *kula or a reflex thereof. 36

Now if, as has been speculated above, Burotu's importance was connected with the trade in red feathers, which was obviously initiated from Polynesia where suitable feathers are not found, then the Pulotu legend relates to a time when the red-feather trade was already established, that is, when Polynesia was already occupied. Moreover, given that the red-feather trade was a frequent cause of wars (Beaglehole 1967:164), the fact that an early movement of people occasioned by a war should originate from an important source of red feathers is not at all remarkable. If, on the other hand, we believe the Pulotu legend to refer to the initial occupation of Polynesia, then we must accept as an extraordinary coincidence the fact that the origin place of the first settlers of Polynesia subsequently became an important source of red feathers, and ultimately the paradise of Fijian mythology.

Another consideration is that, in Anderson's 1777 account from Tonga, it is only the souls of chiefs that go to Pulotu. Although it is stated that all mankind came from Pulotu in Mariner's account, it is again only the souls of chiefs that return there. Collocott (1921:162), likewise, refers to Pulotu as “the home also of spirits of the departed, evidently only of chiefs and great people”. For Samoa also, while Turner (1884:258) states explicitly that all Samoans return to Pulotu after death, most authorities (e.g., Stair 1897:217, Brown 1910:361, 364, and even Turner in an earlier publication 1861:237) agree that the privilege is reserved for chiefs. Thus, Pulotu is probably the origin of not all Tongans and Samoans, but just the more recently arrived stranger-chiefs. For Polynesia generally, Handy (1927:321) concluded from his survey of Polynesian religion that, of the two major immigrant groups, it was the later arrivals that believed that their souls went to an island paradise named Pulotu.

More speculatively, it may be significant that the names of the major Samoan islands mostly have cognates in Eastern Vanualevu or Lau, rather than the Yasayasa Muala or Kadavu area. 37 Except in cases where newcomers outnumber or otherwise overwhelm the earlier population, it seems reasonable to - 366 assume that place-names are more likely to have been determined by the founding population.

Although I have argued that the Pulotu legend does not refer to the initial occupation of Western Polynesia, nevertheless, the events in question probably occurred in the fairly distant past. It has already been noted that cognates of the words pulotu and fanuakula are found throughout Eastern Polynesia, but a notable exception is Rapanui (Easter Island), which appears to lack cognates of both. 38 8 Given that Rapanui is believed to have been settled in the first few centuries A.D. (Kirch 1984:267), the fame of Pulotu would appear to have spread after that time, but before the other extremes of Eastern Polynesia were settled, or during the period after settlement when communication was still maintained, at least indirectly, with Western Polynesia. In this connection, it is pertinent to note that there is increasing archaeological evidence for continuing communication between the West Polynesia-Fiji area and Eastern Polynesia after its settlement, notably the discovery of ceramic sherds of apparent Fijian origin in the Marquesas (Dickinson and Shutler 1974), and of sherds of apparent Tongan origin in Ma'uke in the Cook Islands (Walter and Dickinson 1989).

Referring back to the linguistic arguments for the cognacy of the place-names crucial to this hypothesis: by accepting that the Pulotu incident occurred at a time when Polynesia was already occupied, we may be able to explain the apparently arbitrary split reflexes of Proto Central Pacific *r and *j in Proto Polynesian (and also *c and *z [Geraghty 1986:297-300], which show a similar arbitrary split) as at least partially due to borrowing. Likewise, it may be that there is after all no need to reconstruct PCP *x which, in any case, has no external motivation, since cognates of Fijian/k/ which show PPN *? can be attributed to borrowing from an area of Fiji in which *k had become glottal stop, as is the case in much of eastern Vanualevu today (Geraghty 1984:58). The linguistic scenario suggested is one in which there was extensive borrowing between Fijian and Polynesian languages, and probably also among the languages of Polynesia, long after the Proto Central Pacific languages had begun going their separate ways, so that some lexical items attributable to Proto Central Pacific by the strict comparative method (with the probable exception of those witnessed in Rotuma) may be the product of relatively recent diffusion. I hope to be able to explore the ramifications of this proposal in a future paper.


In the early decades of this century, it was commonly averred by Polynesian ethnologists and historians that there were at least two related, but distinct, major population movements into Polynesia. A number of different names were applied to the two streams, e.g., Polynesians and Maori-Rarotongans (Smith 1910), sitting-interment people (or Proto-Polynesians) and kava people (Rivers - 367 1914:II:479,575), pre-Tangaroans and Tangaroans (Williamson 1924), Indo-Polynesians and Tangaloa-Polynesians (Handy 1927:312); but it was the nomenclature of Proto-Samoans and Tongafiti, as propagated by Smith (1903:3) and Churchill (1908:80, 1911:45), that was most widely used, and referred to in popular literature (Rowe 1930:268, Lambert 1941:109, 167, 184).

According to Smith and Churchill, Samoan traditions tell of a group named the Tongafiti, of unknown origin, who invaded Samoa after its initial occupation and occupied the coasts of Savai'i and 'Upolu, driving the existing Proto-Samoan population into the hills. Smith (1910:164) estimated, on the basis of genealogies, that they were in Samoa for about 550-600 years. Smith (1910:163) originally believed, following Churchill, that it was the Tongafiti who were expelled in the famous incident at Aleipata in which the departing King of Tonga instituted the Samoan title of Mälietoa. Subsequently, however, Smith informed Wilkinson (1924:38) that he believed that this latter incident refers to a later expulsion, not of Tongafiti people, but of Tongans who had been raiding Samoa over a short period; and that the Tongafiti people were earlier invaders of Samoa, and had stayed much longer, than the Tongans. After their expulsion the Tongafiti people fled to the Cook and Austral islands, from which base they spread to the rest of Eastern Polynesia, subjugating, but not linguistically obliterating, the earlier settlers of Proto-Samoan origin (Churchill 1912:44,51).

There is indeed some doubt about the authenticity of the name Togafiti in Samoan history, since it is found in a primary context only in the writings of Smith and Churchill. 39 Although accepted, in whole or in part, by some of their contemporaries(e.g., Rivers 1914:II:584; Williamson 1924:2-3; Handy 1927:328; Haddon and Hornell 1938:73-5), earlier authorities such as Turner (1884), Stair (1897), and Krämer (1902) make no mention of the name, and it is, likewise, absent from modern histories, e.g., Meleisea 1987. Smith (1903:3) and Churchill (1911:45, 180), nevertheless, both state quite clearly that Tongafiti was the name given by the Samoans to their erstwhile enemy.

Despite the lack of corroboration from the literature on Samoan oral traditions, various kinds of evidence do offer some support. First, it should be noted that there is no reason to believe that either Smith or Churchill was in any way dishonest. Smith was the leading Polynesianist of his day, though I can find no reference to his having conducted extensive research in Samoa. He probably obtained his information from Churchill, a leading light in the Polynesian Society of which Smith was founder and president, whom he visited in Samoa in 1897 (Smith 1910:199,258). Churchill served in Apia as Consul-General of the United States from 1896 to 1899, and attributes the bulk of his knowledge of Samoan history to discussions with a “small group of island sages whom I gathered about me in the windward outskirts of Apia”, principal among whom was Lautā, the chief of Vaiala. Churchill also says he spent many hours listening - 368 to Mālietoa Laupepa, the last king of Samoa, of whom he says “few indeed had minds so replete with the myth and tradition in which is preserved the ancient history of his race” (Churchill 1911:iv). Unfortunately, no results of this historical research appear to have been published. Presumably they are contained in the manuscript entitled “Samoa o le Vavau” which was, in 1911, “awaiting its due season” (Churchill 1911:387). The impression gained from his voluminous linguistic work is that, while not slow to point out his own achievements, he was an innovative thinker and a meticulous scholar.

Second, there is comparative linguistic evidence. In Kiribati, according to Sabatier's (1971) dictionary, the word tongabiti means “a tribe driven out of Samoa about XIII century, scattered from Tonga to Gilbert Islands”. The date and other information given suggest that the definition was at least influenced by knowledge of Smith's and Churchill's theories, but there is no direct indication that the word is a recent introduction. Against this must be set the fact that the word is missing in detailed accounts of Kiribati oral tradition: although the Tongafiti story is noted in Grimble 1972:56, it is not given as a Kiribati tradition. On the contrary, Grimble (1933:81, 82; 1972:273; 1989:293) states that the Samoan invaders called themselves “the seed of Matang”, “the breed of red men”, “the breed of Samoa”, and various clan-specific names, but not tongabiti. So the evidence from Kiribati is somewhat inconclusive.

Somewhat more encouraging, however, is the fact that reflexes of the word tongafiti are found all over Polynesia with meanings such as “device, stratagem”, “foreigner, enemy”, and “chief” that could conceivably be derived from the name of an invading people who gained supremacy and were noted for their skill and cunning. 40 Moreover, the Marquesan cognate Tona-fiti is said to be the name of the chief of the Marquesan people when they resided in their ancient homeland Hawai'i (Smith 1910:118). The Rarotongan cognate To'ahiti is the name of the god who is said to have come from 'Avaiki and discovered Rarotonga (Te-aia 1893; Smith 1910:231, 279; Te Ariki-tara-are 1920:1-2; Buck 1938:114), and is associated in myth with the god Tangaroa (Stair 1895b:103,129). In New Zealand Maori oral history, Tongawhiti is the name of an island sometimes mentioned in old Maori poetry, which Smith (1891:306) takes to refer to one of the Fiji islands, and Tonga-whiti-atea was the name of a man of Hawaiki whom Pou met when he went there to procure the kūmara (sweet potato) (Anon. 1921).

The relevance of this discussion of the name Tongafiti is that there is in Matuku, and nowhere else in Fiji to my knowledge, a place named Togaviti. 41 It is a prominent hill behind Makadru, approximately 1.5 km to the south-east, and the highest point in the southern half of Matuku, said to have been occupied in immediate pre-Christian times by one or more of the clans of Makadru. It is possible that Togaviti in Matuku was the original home of the Tongafiti people, - 369 and that the Pulotu legends refer to this population movement, or a related incident.

Apart from the Matuku connection, there are other similarities between the Pulotu and Tongafiti stories. The god Tangaloa who, in Tongan legend, ordered the move from Pulotu to Tonga and who, in one version of the Samoan legend, is the father of the two rulers of Pulotu, is associated in Rarotonga with the god To'ahiti. The name of one of those involved in the Pulotu war, by Turner's account, was Falï. In Smith's (1910:158-60) account of the war which caused the Tongafiti people to spread out from Fiji to Polynesia, based on Rarotongan oral history, the name of one of the three people said to have originated the war is Ari. Unfortunately, Smith is inconsistent in the marking of both glottal stop and vowel length, so it is not possible to state whether these two forms are perfectly cognate. More valuable perhaps, because it may be archaeologically testable, is the connection with war. The Pulotu incident related by Turner was said to be “the beginning of wars”, and the Tongafiti people were warlike invaders who succeeded in driving the “Proto-Samoans” into the interior of the two main islands of Samoa. Moreover, an initial comparison of dates is encouraging: the Tu'i Tonga line, source of all Tongan royalty so putatively of Pulotu origin, is reckoned genealogically to have emerged about A.D. 1,000 (Poulsen 1977:23), and approximately the same date is assigned to the sudden appearance of large hill-forts in Lau (Best 1984:644), indicating the onset of warfare. 42 It is probably also significant that Samoan adzes appear in Lakeba at about this time (Best 1984:644). If the prehistoric kula trade was like the historic, then it was the Tongans who were the middlemen, buying in Fiji and selling in Samoa (Clunie 1986:150-1), and the goods acquired by Fijians in the transaction would have been Tongan goods, which are mostly archaeologically invisible (Kaeppler 1978:252), hence the lack of archaeological evidence for exchange with Western Polynesia in the first millennium A.D. The move of the Tongafiti people to Samoa would have revived the direct link between Fiji and Samoa, and facilitated the supply of the superior Samoan adzes to Fiji.

If it is indeed the case that the Pulotu and Tongafiti incidents are one and the same, and we accept the linguistic and mythological evidence that the Tongafiti eventually spread to all of triangle Polynesia except Rapanui, then it follows that the immediate extra-Polynesian origin of at least some of the Rarotongan and New Zealand Maoris is in the Matuku area. So that, just over 100 years after the founding of the Society which has devoted much of its energies to investigating the “whence of the Maori”, it may be that a plausible, if partial, solution has been found.

- 370

The following is a brief summary of the sequence of events proposed, tentatively or otherwise, in this paper.

  • 1. After the initial settlement of Western Polynesia from Eastern Fiji, a dialect chain develops and phonological differences increase, until the Proto Polynesian sound system is distinct from that of Eastern Fijian.
  • 2. Polynesians, probably principally or exclusively Tongans, come to Fiji to barter for kula (red feathers) and find a good supply, including both collared lory and Kadavu parrot, in Eastern Lau.
  • 3. After the supply of kula is depleted in Eastern Lau, perhaps about 2,000 B.P., Matuku becomes the centre for the trade, and with the wealth and prestige thus acquired becomes also a centre of culture. Under the name Fanuakula (Golden Land) its fame spreads throughout Polynesia, with the exception of the now isolated Rapanui.
  • 4. Competition for dwindling kula resources arises between two halves of Matuku, Burotu and Babajea, resulting in war, and people moving to Polynesia, some perhaps by way of Lakeba, about 1,000 B.P. These people, more skilled in warfare, overcome the coastal Samoans and drive them into the hills, occupying parts of Samoa for maybe 500 years. They are known as the Tongafiti people, after their main hill-fort in Matuku. In Tonga they become the ruling chiefs. They seem to have then spread through much of Polynesia, again excepting Rapanui, and to have had a major impact on many of the existing populations, the precise nature of which remains to be explored.
  • 5. Matuku declines with the depletion of its kula resources, and the trade moves to Kadavu and more distant islands such as Vanuatu and Pohnpei. Burotu becomes equated in Fijian myth with the disappearing island peopled only by females, full of golden produce, and source of all things exotic, its mystique possibly enhanced by a natural catastrophe of some description. In Polynesia, the exact location of the Tongafiti homeland Pulotu is soon forgotten, and it becomes an abode of departed spirits and source of exotic produce.

An important lesson to be learned from this paper is the value of the identification of loan phonology. Linguists have long been in possession of this very valuable tool in reconstructing prehistory, but have failed to apply it to any effect. It is not infallible, but the conditions of extensive bilingualism in languages with transparent sound correspondences that are required to vitiate it are quite rare. Had such a tool been available at the time, Smith (1894:145) - 371 would not have talked of the “large number of purely Polynesian words incorporated in the Fijian language” or the “evident Polynesian origin of many of the place-names in the Fijian group, especially those of the eastern part”, a view that was widely held (e.g., by Fornander 1878:I:33; Churchill 1908:150). 43 I have used this tool to identify introduced plants in Fiji and Polynesia (Geraghty 1990:89) and again in this paper to demonstrate that Fijian Burotu is probably not a loan from Polynesian Pulotu, whereas Fijian kula probably is a Polynesian loan.

Another case in point is the naming of the Polynesian Outliers. Many have speculated (e.g., Hale 1846:186-7) on the significance of the similarity of names between Tikopia, a Polynesian Outlier in the Eastern Solomons, and Cikobia, a remote island at the north-east extremity of the Fiji group. As noted by Hocart (1952:4), the correspondences indicate a loanword, i.e., that Tikopia is borrowed from Fijian Cikobia. 44 So it is unlikely that Tikopia was named during the first thousand years or so of the occupation of west Polynesia, when it formed a dialect chain with Fiji. The most likely explanation is that the Polynesians who named Tikopia did so because they believed they were on Cikobia, or perhaps could not conceive of where else they might be. Similarly, those who named Tikopia's neighbour Anuta, which is also a loanword, believed they were on Yanuca, 45 another eastern outlier of the Fiji group, some 30 km north-east of Taveuni. If we accept that such naming occurred, the problem of the outliers Uvea, Futuna, and Aniwa (Niua) being not particularly close linguistically to their namesakes disappears: they were not named in remembrance of the homeland, but were simply a case of misidentification. 46

Another lesson is that there is much of value to be gleaned from the works of Turner, Smith, Stair, Churchill, and their like. They were primarily collectors and translators of oral traditions, as was the express aim of the Polynesian Society (Smith 1910:257), and they did an admirable job of salvage ethnology under difficult circumstances. If we have found that we cannot subscribe to their more fanciful theories — that the Polynesians originated from India, that they migrated en masse from Indonesia to Polynesia, and so on — it is because we have access to more and better data, and science has progressed. Much of what they recorded remains extremely valuable, and sadly underused.

- 372 Page of endnotes

- 373 Page of endnotes

- 374 Page of endnotes

- 375 Page of endnotes

- 376 Page of endnotes

- 377 Page of endnotes

- 378
  • Andersen, Johannes C., 1928. Myths and Legends of the Polynesians. London: Harrap.
  • Anderson, William, 1967 [1776-7]. A Journal of a voyage made in His Majestys Sloop Resolution, in J. C. Beaglehole (ed.), The Voyage of the Resolution and Discovery 1776-1780. Hakluyt Society Extra Series XXXVI. Cambridge: Cambridge University Press, pp. 723-986.
  • Anon., 1892. Genealogies and historical notes from Rarotonga. Journal of the Polynesian Society, 1:20-9.
  • —— 1921. The visit of Pou to Hawaiki to Procure the kumara. Journal of the Polynesian Society, 30:40-4.
  • Barker, George T., 1925a. Vanua Alewa, the Sirens' Isle of Fiji. Transactions of the Fijian Society for the year 1924, 25-9.
  • —— 1925b. Dakuwaqa. Transactions of the Fijian Society for the year 1924, 30-6.
  • —— 1927. Conua, a Town on the Nakauvadra. Transactions of the Fijian Society for the years 1925 and 1926, 1-5.
  • Beaglehole, Ernest and Pearl, 1938. Ethnology of Pukapuka. Bernice P. Bishop Museum Bulletin 150. Honolulu: Bishop Museum.
  • Beaglehole, J. C. (ed), 1967. The Journals of Captain James Cook on his Voyages of Discovery. The Voyage of the Resolution and Discovery 1776-1780. Published for the Hakluyt Society. Cambridge: Cambridge University Press.
  • —— 1968. The Journals of Captain James Cook on his Voyages of Discovery. The Voyage of the Endeavour 1768-1771. Published for the Hakluyt Society. Cambridge: Cambridge University Press.
  • Beckon, William N., MS 1989. Distribution of Land Birds in Fiji: a Provisional Annotated list, by Island. Department of Zoology, University of California at Davis.
  • Besnier, Niko, 1981. Ttou Tauloto te Ggana Tuuvalu: a Course in the Tuvaluan Language. Funafuti: United States Peace Corps.
  • Best, Simon B., 1984. Lakeba: the Prehistory of a Fijian Island. Unpublished Ph.D.thesis, University of Auckland
  • ——, Peter Sheppard, Roger Green and Robin Parker, 1992. Necromancing the Stone: Archaeologists and Adzes in Samoa. Journal of the Polynesian Society, 101,1:45-85.
  • Biggs, Bruce, 1990. The Comparative Polynesian Lexicon Project. Computer file. Department of Anthropology, University of Auckland.
  • Blust, Robert A., 1976. A Third Palatal Reflex in Polynesian Languages. Journal of the Polynesian Society, 85:339-58.
  • —— 1978. The Proto-Oceanic Palatals. Polynesian Society Memoir No. 43. Wellington: Polynesian Society.
- 379
  • Brown, George, 1910. Melanesians and Polynesians. Their Life-histories Described and Compared. London: Macmillan.
  • Buck, Peter H., 1938. Vikings of the Sunrise. Philadelphia: Lippincott. Burrows, Edwin G., 1936. Ethnology of Futuna. Bulletin 138. Honolulu: B. P. Bishop Museum.
  • —— 1937. Ethnology of Uvea (Wallis Island). Bulletin 145. Honolulu: B. P. Bishop Museum.
  • —— 1938. Western Polynesia: A Study of Cultural Differentiation. Etnologiska Studier. 7:1-192.
  • Cagileta, Kolinio, MS n.d. [1938]. Tukutuku Raraba, Yavusa Qalikarua. Typescript. Native Lands Commission, Suva.
  • Capell, A., 1941. A New Fijian Dictionary. Sydney: Australasian Medical Publishing Company.
  • Carrol, A., 1895. Location of Bulutu, Buruttu, or Pulotu. Journal of the Polynesian Society, 4:153-4.
  • Churchill, William, 1908. Samoan Phonetics in the Broader Relation. Journal of the Polynesian Society, 17:79-92, 149-61, 209-21.
  • —— 1911. The Polynesian Wanderings. Washington: Carnegie Institution.
  • —— 1912. Easter Island: the Rapanui Speech and the Peopling of South-east Polynesia.Washington: Carnegie Institution.
  • Churchward, C. M., 1940. Rotuman Grammar and Dictionary. Sydney: Australasian Medical Publishing Company.
  • Clunie, Fergus, 1984. Birds of the Fiji Bush. Suva: Fiji Museum.
  • —— 1986. Yalo i Viti: A Fiji Museum.Catalogue. Suva: Fiji Museum.
  • Codrington, R. H., 1891. The Melanesians. Oxford: Clarendon Press.
  • Cokanasiga, Alifereti, 1990. Na kau vakavesu nei Ravouvou kei Vanuakula. Tape recording and transcript, Institute of Fijian Language and Culture, Suva.
  • Collocott, E. E. V., 1921. Notes on Tongan Religion. Journal of the Polynesian Society 30:152-63,227-40.
  • Coriakula, Seremaia, MS n.d. [1938]. Tukutuku Raraba, Yavusa Levukaidaku. Typescript. Native Lands Commission, Suva.
  • Dening, G. M., 1962. The Geographical Knowledge of the Polynesians and the Nature of Inter-island Contact, in Jack Golson (ed.), Polynesian Navigation: a Symposium on Andrew Sharp's Theory of Accidental Voyages. Wellington: Polynesian Society, pp. 102-53.
  • Dickinson, W. R. and R. Shutler, Jr, 1974. Probable Fijian Origin of Quartzose Temper Sands in Prehistoric Pottery from Tonga and the Marquesas. Science, 185:454-7.
  • Dixon, Roland B., 1916. The Mythology of All Races: Oceanic. Boston: Marshall Jones.
  • Eason, W.J. E., n.d. [1951]. A Short History of Rotuma. Suva: Government Printing Department.
  • Elbert, Samuel H., 1975. Dictionary of the Language of Rennell and Bellona. Copenhagen: National Museum of Denmark.
  • Englert, Sebastian, 1978. Idioma Rapanui. Santiago de Chile: Ediciones de la Universidad de Chile.
- 380
  • Firth, Raymond, 1985. Tikopia-English Dictionary. Auckland: Auckland University Press and Oxford University Press.
  • Fison, Lorimer, 1907. Tales from Old Fiji. London: De La More Press.
  • Fornander, Abraham, 1878. An Account of the Polynesian Race. London: Trübner.
  • France, Peter, 1969. The Charter of the Land: Custom and Colonization in Fiji. Melbourne: Oxford University Press.
  • Geraghty, Paul, 1983. The History of the Fijian Languages. Oceanic Linguistics Special Publication 19. Honolulu: University of Hawaii Press.
  • —— 1986. The Sound System of Proto-Central-Pacific, in Paul Geraghty, Lois Carrington and S. A. Wurm (eds), Focal II: Papers from the Fourth International Conference on Austronesian Linguistics. Pacific Linguistics C94. Canberra: Australian National University, pp. 289-312.
  • —— 1989. Linguistic Evidence for the Tongan Empire. Paper presented at the Comparative Austronesian Project Symposium on Contact-induced Language Change, 26th August - 1st September 1989, at the Department of Linguistics, Research School of Pacific Studies, Australian National University.
  • —— 1990. Proto-Eastern Oceanic *R and its Reflexes, in J. H. C. S. Davidson (ed.), Pacific Island Languages: Essays in Honour of G. B. Milner. London: School of Oriental and African Studies, University of London, pp. 51-93.
  • —— in press. Linguistic Evidence for the Tongan Empire, in Tom Dutton and Darrell Tryon (eds), Language Contact and Change in the Austronesian World. Mouton de Grutyer.
  • Gibbons, John R. H., and Fergus G. A. Clunie, 1986. Sea Level Changes and Pacific Prehistory. Journal of Pacific History, 21:58-82.
  • Gifford, Edward Winslow, 1924. Tongan Myths and Tales. Bulletin 8. Honolulu: B. P. Bishop Museum.
  • Gill, W. Wyatt, 1876. Myths and Songs from the South Pacific. London: King.
  • —— n.d. [1876]. Life in the Southern Isles. London: Religious Tract Society.
  • Goodenough, Ward H., 1961. Migrations Implied by Relationships of New Britain Dialects to Central Pacific Languages. Journal of the Polynesian Society, 70:112-26.
  • —— 1986. Sky World and This World: the Place of Kachaw in Micronesian Cosmology. American Anthropologist, 88:551-68.
  • Grimble, Arthur, 1933. The Migrations of a Pandanus People. Memoir No. 12 (uncompleted). Journal of the Polynesian Society supplement 42:1-50, 51-84.
  • —— 1972. Migrations, Myth and Magic from the Gilbert Islands. Arranged and illustrated by Rosemary Grimble. London and Boston: Routledge and Kegan Paul.
  • —— 1989. Tungaru Traditions: Writings on the Atoll Culture of the Gilbert Islands. Edited by H. E. Maude. Pacific Islands Monograph series, No. 7. Honolulu: Center for Pacific Islands Studies, University of Hawaii.
  • Haddon, A. C. and James Hornell, 1938. Canoes of Oceania. Volume III. Definition of Terms, General Survey, and Conclusions. Special Publication 29. Honolulu: B. P Bishop Museum.
  • Hale, Horatio, 1846. United States Exploring Expedition: Ethnography and Philology. Philadelphia: Lee and Blanchard.
- 381
  • Handy, E. S. Craighill, 1927. Polynesian Religion. Bulletin 34. Honolulu: Bishop Museum.
  • Heatley, E. M., 1923. The Invisible Isle of Burotu. Transactions of the Fiji Society for the year 1922.
  • Herman, [Brother], 1955. Tala o le Vavau: Samoan Legends Collected and Translated by Brother Herman. Pago Pago: Association of the Marist Brothers' Old Boys.
  • Hocart, A. M., 1929. Lau Islands, Fiji. Bulletin 62. Honolulu: B. P. Bishop Museum.
  • ——1952. The Northern States of Fiji. Occasional Publication No. 11. London: Royal Anthropological Institute.
  • Hooper, Robin, in press. Reconstructing Proto-Polynesian Fish Names, in Malcolm Ross and Andrew Pawley (eds), Austronesian Terminologies: Continuity and Change. Canberra: Pacific Linguistics.
  • Hovdhaugen, Even, 1985. A Sketch of Samoan Grammar. Oslo, published by the author.
  • ——1987. From the Land of Nāfanua. Oslo: Norwegian University Press and the Institute of Comparative Research in Human Culture.
  • Ivens, W. G., 1927. Melanesians of the South-east Solomon Islands. London: Kegan Paul, Trench, Trubner.
  • Jackson, Frederick H., 1986. On Determining the External Relationships of the Micronesian Languages, in Paul Geraghty, Lois Carrington and S. A. Wurm (eds), Focal II: Papers from the Fourth International Conference on Austronesian Linguistics. Pacific Linguistics C94. Canberra: Australian National University, pp. 201-38.
  • Kaeppler, Adrienne L., 1978. Exchange Patterns in Goods and Spouses: Fiji, Tonga and Samoa. Mankind, 11:246-52.
  • Kirch, Patrick Vinton, 1984. The Evolution of the Polynesian Chiefdoms. Cambridge: Cambridge University Press.
  • ——1988. Niuatoputapu: the Prehistory of a Polynesian Chiefdom. Thomas Burke Memorial Washington State Museum Monograph No. 5. Seattle: Burke Museum.
  • Krämer, Augustin, 1902. Die Samoan-Inseln. Vol. 1. Stuttgart: Schweizerbartsche Verlagsbuchhandlung.
  • Lambert, S. M., 1941. A Yankee Doctor in Paradise. Boston: Little, Brown.
  • Levy, Richard, n.d. Languages of the Southeast Solomon Islands and the Reconstruction of Proto Eastern Oceanic. Unpublished paper, Department of Anthropology, University of Kentucky.
  • McEwen, J. M., 1970. Niue Dictionary. Wellington: Department of Maori and Island Affairs.
  • Malinowski, B., 1929. The Sexual Life of Savages in North-Western Melanesia. New York: Halcyon.
  • Martin, John, 1818. An Account of the Natives of the Tonga Islands. London: Murray.
  • Mead, Margaret, 1969. Social Organization of Manu'a. Bulletin 76. Honolulu: Bishop Museum Press.
  • Meleisea, Malama, 1987. Lagaga: a Short History of Western Samoa. Suva: University of the South Pacific.
  • Milke, Wilhelm, 1968. Proto-Oceanic Addenda. Oceanic Linguistics, 7:147-71.
- 382
  • Milner, G. B., 1966. Samoan Dictionary. London: Oxford University Press.
  • Mocevakaca, Isimeli, MS n.d. [1938]. Tukutuku Raraba, Yavusa Yaroi. Typescript. Native Lands Commission, Suva.
  • Mokunitulevu Na Rai [pseud. for Epeli Rokowaqa], n.d. [1928?]. Ai Tukutuku kei Viti. [Suva: Methodist Church in Fiji].
  • Moseley, H. N., 1944 [1879]. Notes by a Naturalist made during the Voyage of H.M.S. “Challenger”. London: Laurie.
  • Moyle, Richard M., 1988. Traditional Samoan Music. Auckland: Auckland University Press in Association with The Institute for Polynesian Studies.
  • Newell, J. E., 1895. Notes on Rev. J. B. Stair's Paper on “Early Samoan voyages and settlement”. Journal of the Polynesian Society, 4:238-40.
  • Niu Madu [pseud.], 1894. Ai Tukutuku Eso mai na Veiyasana: Lau. Na Mata, 47:164-5
  • Nunn, Patrick D., 1987. Sea-level and Tectonic Changes in the Pacific Basin, with Emphasis on Holocene Changes in Fiji. Domodomo, 5:4-15.
  • —— 1990. Coastal Processes and Landforms of Fiji: their Bearing on Holocene Sealevel Changes in the South and West Pacific. Journal of Coastal Research, 6:279-310.
  • Pétard, Paul, 1986. Quelques Plantes Utiles de Polynésie Française et Raau Tahiti. Tahiti: Éditions Haere Po no Tahiti.
  • Pritchard, W. T., 1968 [1866]. Polynesian Reminiscences. London: Dawsons.
  • Pukui, Mary Kawena and Samuel H. Elbert, 1986. Hawaiian Dictionary: Hawaiian-English, English-Hawaiian. Revised and enlarged edit. Honolulu: University of Hawaii Press.
  • Qio, Sepeti, MS n.d. [1938]. Tukutuku Raraba, Yavusa Lomati. Typescript. Native Lands Commission, Suva.
  • Reid, A. C, 1990. Tovata I & II. Suva: Fiji Museum.
  • Rinke, Dieter, 1989. The Relationship and Taxonomy of the Fijian Parrot Genus Prosopeia. Bulletin of the British Ornithologists' Club, 109:185-95.
  • Rivers, W. H. R., 1914. The History of Melanesian Society. Cambridge: Cambridge University Press.
  • Rochas, Victor de, 1859. Anthropologie (de la Nouvelle-Calédonie). Rev. Algérienne et Coloniale, 1:224-35.
  • Rowe, N. A., 1930. Samoa under the Sailing Gods. London and New York: Putnam.
  • Sabatier.E., 1971. Gilbertese-English Dictionary. [?] Noumea: South Pacific Commission Publications Bureau.
  • Sahlins, Marshall D., 1962. Moala: Culture and Nature on a Fijian Island. Ann Arbor: University of Michigan Press.
  • St. Johnston, T. R., 1918. The Lau Islands (Fiji) and Their Fairy Tales and Folk-lore. London: Times Book Co.
  • —— 1921. The Islanders of the Pacific. New York: Appleton.
  • Schultz, E., 1980. Samoan Proverbial Expressions. Auckland: Polynesian Press and Suva: Institute of Pacific Studies, University of the South Pacific.
  • Smith, S. Percy, 1891. Notes on the Geographical Knowledge of the Polynesians, Part
- 383
  • I. Paper read before the Australasian Association for the Advancement of Science, January 1891.
  • ——1894. The Polynesian Sojourn in Fiji. Journal of the Polynesian Society, 3:145-52.
  • —— 1898. Notes on the Geographical Knowledge of the Polynesians, Part II. Paper read before the Australasian Association for the Advancement of Science, January 1898.
  • ——1903. Niuē Island, and its People. Part IV. History and Traditions. Journal of the Polynesian Society, 12:1-31.
  • —— 1910. Hawaiki: the Original Home of the Maori. Auckland: Whitcombe and Tombs.
  • Stair, J. B., 1895a. Samoa: Whence Peopled?Journal of the Polynesian Society, 4:47-58.
  • ——1895b. Floatsam [sic] and Jetsam from the Great Ocean: or, Summary of Early Samoan Voyages and Settlement. Journal of the Polynesian Society, 4:99-131.
  • —— 1897. Old Samoa, or, Flotsam and Jetsam from the Pacific Ocean. Oxford: Religious Tract Society.
  • Stimson, J. Frank, 1932. Songs of the Polynesian Voyagers. Journal of the Polynesian Society, 41:81-201.
  • Sukanasau, Meli, 1979. Ko Tui Liku. Tape recording and transcript, Institute of Fijian Language and Culture, Suva.
  • Tagilala, Josefa, MS 1989. Na Noqu Koro. Unpublished manuscript, Institute of Fijian Language and Culture, Suva.
  • Te-aia, 1893. Genealogies and Historical Notes from Rarotonga. Part III. Journal of the Polynesian Society, 2:271-8.
  • Te Ariki-tara-are, 1920. History and Traditions of Rarotonga. Part VII. Translated by S. Percy Smith. Journal of the Polynesian Society, 29:1-11.
  • Thompson, Laura, 1940. Southern Lau, Fiji, an Ethnography. Bulletin 162. Honolulu: B. P. Bishop Museum.
  • Thomson, Basil, 1892. The Land of our Origin (Viti, or Fiji). Journal of the Polynesian Society, 1:143-6.
  • ——1908. The Fijians: a Study of the Decay of Custom. London: Heinemann.
  • Tikoisuva, Netani, MS n.d. [1938]. Tukutuku Raraba, Yavusa Dakuimatuku. Typescript. Native Lands Commission, Suva.
  • Tregear, Edward, 1891. The Maori-Polynesian Comparative Dictionary. Wellington: Lyon and Blair.
  • Tua, Benjamin, 1979. Our Custom Way Vanikoro. Edited by Peter C. Lincoln. Auckland: University of Auckland Bindery.
  • Tuinaceva, Wiliame, MS n.d. [1939]. Tukutuku Raraba, Yavusa Delaikorolevu. Typescript. Native Lands Commission, Suva.
  • Tuiteleleapaga, Napoleone A., 1980. Samoa Yesterday, Today and Tomorrow. Great Neck, New York: Todd and Honeywell.
  • Turner, George, 1861. Nineteen Years in Polynesia. London: Snow.
  • ——1884. Samoa. London: Macsmillan.
- 384
  • Vakatovolea [pseud.], 1896. Ai Talanoa Makawa. Na Mata, 68:124.
  • >Walter, Richard and W. R. Dickinson. A Ceramic Sherd from Ma'uke in the Southern Cook Islands. Journal of the Polynesian Society, 98:465-70.
  • Waterhouse, Joseph, 1866. The King and People of Fiji.London: Wesleyan Conference Office.
  • Watling, Dick, 1982. Birds of Fiji, Tonga and Samoa. Wellington: Millwood.
  • Williams, John, 1842. A Narrative of Missionary Enterprises in the South Sea Islands. London: Snow.
  • Williams, Thomas, 1870. Fiji and the Fijians. London: Hodder and Stoughton.
  • Williamson, R. W., 1924. The Social and Political Systems of Central Polynesia. Cambridge: Cambridge University Press.

Caginavanua, Ilaitia. Former Acting Commissioner of Native Lands and Fisheries, Fiji; Helu, Futa. Director, 'Atenisi Institute, Tonga; Langdon, Robert. Pacific historian, Australian National University; Nunn, Patrick. Senior Lecturer in Geography, University of the South Pacific; Sogotubu, Rusiate. Late Roko Sau (High Chief) of Totoya; Soko, Setareki. Retired Native Magistrate, Nukuni, Ono.

1   This paper is a substantially revised and expanded version of a paper entitled “Proto Central Pacific, Pulotu, and the Polynesian Homeland” presented at the Sixth Internatinal Conference on Austronesian Linguistics in Honolulu, Hawaii, in 1991. My thanks to David Routledge, Patrick Nunn, Randy Thaman, Jeffrey T. Clark, Robert Langdon, Futa Helu, Dick Watling Marshall Sahlins and John Lynch for helpful comments on an earlier draft; to Setareki Tuinaceva, Margaret Patel, and the Staff of the National Archives of Fiji, Inoke Qalo, Pio Mōnoa, Ann Chowning, Peter Lincoln, and Even Hovdhaugen and Ross Clark for help with references; to Viliame Sisikakala for the original map; and especially to Andrew Crosby, for manysuggestions and references while this paper was being extensively revised. None of these kind people is to be held responsible for any of the views expressed in this paper. Unless otherwise indicated, data are based on interviews with many Fijians, especially from Matuku and Ono.
2   Hale (1846:195) proposed Buru, in eastern Indonesia, as a possibility, no doubt swayed by the widespread belief that the Polynesians must have bypassed or skirted Melanesia on their way east. Even apart from the geographical distance, the formal problems with this proposed cognate pair are enormous. Among the more remarkable locations proposed since, mention must be made of Carroll's (1895:154) assertion that Pulotu in fact refers to “Burattu, or Burutu, along the central part of the Euphrates River, in Mesopotamia”
3   The following words are, however, probably related: Tikopia purotu “waterspout, traditionally regarded as a spirit manifestation” (Firth 1985), Niue pulotu “wise” (McEwen 1970), Tikopia purotu “person of skill and judgement”, and the widespread Eastern Polynesian purotu “beautiful” (Tregear 1891). Andersen (1928:420) has pointed out a possible connection between Maori purotu “pure, clean, transparent” and the lake of eternal youth (vaiola believed to be situated in Pulotu. Note also that pulotu is found in many Western Polynesian and Outlier languages with the meaning “teacher of songs and dances”, and in Samoan is a kind of drum reserved for chiefs (Stair 1897:135; Moyle 1988:41).
4   Note the conflicting etymologies for ‘Upolu: the island-naming version is probably of recent origin. Certainly it appears to be the case in Fiji that the art of deriving place-names from the sayings and doings of ancestor heroes (deriving them from the compounding of spouses’ names is practically unknown) is largely recent development, which blossomed with the establishment of the Native Lands Commission a century or so ago, since word quickly spread that tribal histories containing such etymologies were more likely to be credited. Mead (1969:156) seems to share this scepticism: “This punning mythology is of never-failing interest to the Samoans: endless incompatible tales are built up by splitting up place names and family names and fabricating myths about the linguistic elements”. Turner (1884:10-6) gives four different popular etymologies of the name Samoa.
5   This may, however, be due to common inheritance of an older belief: in Trukese cosmology, a goddess described as half-moray-eel and half-woman tries to catch and eat spirits attempting to cross the sea separating the human world from the region of the gods (Goodenough 1986:557).
6   This statement requires some qualification. Although the current standard myth about a fleet sailing from Tanganyika is clearly a European-inspired invention of the late 19th century, early records of oral tradition (Hale 1846:52; Thomson 1892:145; Barker 1927:1) do indicate that either Degei, the supreme deity and ancestor of many chiefs of Eastern Fiji, or an ancestor of his, first appeared as a stranger on the Western coast of Vitilevu.
7   North-east Vanualevu bulu, a locative noun meaning “under the earth, under the water”, and bulu-t, a widespread verb meaning “bury”, may be etymologically related to this mythological place-name.
8   References to Burotu as an abode of departed spirits do occur (e.g., Pritchard 1968 [1866]:364, 382; Thomson 1908:117; St. Johnston 1918:29), but are explicable as assumptions based on familiarity with Polynesian Pulotu, in much the same way as Fijian mana has been misinterpreted by generations of writers assuming it has the same meaning as Polynesian mana. Thompson (1940:115) and Niu Madu (1894:165) both report that the people of Ono say their souls go to Burotu. If these reports are true, Ono is unique in this respect; but it must be remembered that Ono is the closest Fijian island to Tonga, and the oral traditions of the Ono people say the original inhabitants were Tongans.
9   Burotu is the name of a village in Veinuqa (Namosi), a piece of land on the coast of Bāravi (Nadrogā), and kin-groups in Vadravadra (Bulu, Bā), Nāmata (Tailevu), Nadoi (Rewa), and Tokalau (Kabara, Lau) (Caginavanua, personal communication). As far as I have been able to determine, no particular significance is attached to the name in any of these places.
10   Muala is the original Fijian pronunciation. The name Moala found on most maps and in the literature, and currently widely used in Fiji, is derived from the Tongan form. A number of other Tonganised place-names have come into general use in favour of their Fijian originals, one frequently encountered being Fulaga, for Vulaga.
11   Kubunavanua is the Standard Fijian form of his name. In the Lau islands, which include Matuku, he is known as Kubuavanua.
12   Some of these sightings may be of pumice floating from Tonga seen from a distance (P. Nunn, personal communication). The same desire to distinguish Burotu from drifting pumice may be the reason Tongans occasionally refer to Pulotu as Pulotu Tu'umau (permanent Pulotu).
13   The miji (Polynesian starling, Aplonis tabuensis) is a largish bird and does not flit among the coconut trees. This may be a reference to the kula (blue-crowned lory, Vini australis) which streaks rather than flits among coconut trees, but does have the Burotu-esque “bright red cheeks, chin, throat and upper breast, and a red belly patch” (Clunie 1984:60); but the most likely candidate is the orange-breasted honeyeater (Myzomela jugularis), which does flit among coconut trees, and the male of which boasts a red rump and very conspicuous red nape (Clunie 1984:114). The Lakeba name, which St. Johnston would have been most familiar with, since Lakeba is the administrative centre of Southern Lau, is mijikula. In Ono it is the bicibicikula.
14   No one I have asked has been able to identify Nasali. Vakatovolea (1896) states that it was the name given to Tuvana by the lewa matagi, but the words of the rhyme hardly support this claim. Fison (1907:16) identifies Nasali as “a little island which you may see from the mast of a canoe in the Ono passage” but, apart from the Tuvanas, no island fits this description.
15   Beaglehole does not disclose the source of this information, but one possibility is Teuira Henry's Ancient Tahiti(Langdon, personal communication). Smith(1898:5) also glosses Papatea in an ancient Tahitian chant as Ma'atea; the context suggests that Teuira Henry was also the source of this identification. The chant states that Papatea is the place of origin of the god Hau.
16   Papatea is also the name of a place on Ta'ū, American Samoa, of which Tuiteleleapaga (1980:15) gives the following account: “Those couples or families who had survived the depopulation of Papatea by King Elo and his men, by escaping in their canoes, called on the various atolls or islands throughout the Pacific and inhabited them. One couple, Ma and his wife Nu'a, fearing that the tyrant king might overcome them in his pursuit, kept on sailing until they arrived at what is now Ta'ū, Manu'a. They found a cave on the farthest end of the present village, to the west, which they thought was safe. There they stayed and called the place “Papatea” in memory of their beloved home”. It is presumably this Papatea that was recorded by Anderson (as Pappataia, which corresponds perfectly according to Anderson's phonetic system) during Cook's 1777 visit to Tonga (Cook and King 1784:369; Geraghty, in press). It was said by Tongans to be a large island and was listed after Tutuila, Manuka, and Olosega, so appears to have been a general name for the island of Ta'ū, for the Tongans at least.
17   In Fiji, honorific names are usually names of old village sites, though no one I have asked is able to attribute a specific location to Babasea.
18   This identification was first proposed by Smith (1910:114).
19   The reconstruction of POC *kadavu “raincloud” is based on the following data: Bulu (New Britain) kandavu “rain” (Goodenough 1961:116)
Proto East Admiralties *kandahV “cloud” (Blust, personal communication in Jackson 1986:226)
Proto Micronesian *kaja(w)u “cloud, raincloud” (Jackson 1986:226)
Proto North Vanuatu *kadau, kadua (Mota xanue “west wind”, Raga xadue “southwest wind”; parallel raising of *a is found in Mota vunue, Raga venue from PNV *venua “k tree, Macaranga”)
The Samoan word atafu “exposed to the sun” (Milner 1966) may be related, but the meaning suggests rather a relationship with tafu “be alight, burn”, which has a derivative fa'atafu “in the heat of the day”. I have been unable to find any other possible cognates in Central Pacific languages. A cognate of this Samoan word may well be the source of the name of the island Atafu in Tokelau.
20   In presenting linguistic data, the following language abbreviations are used:
EP Eastern Polynesian, EUV East Uvea, KAP Kapingamarangi, MAE Mae, Vanuatu, MQN Northern Marquesas, NUK Nukuoro, PCP Proto Central Pacific, PEO Proto Eastern Oceanic, PFJ Proto Fijian, PMC Proto Micronesian, POC Proto Oceanic, PPN Proto Polynesian, PSS Proto Southeast Solomons, PTK Proto Trukic, REN Rennell, ROT Rotuma, SAM Samoa, TIK Tikopia, TON Tonga, TVL Tuvalu. For sources and other conventions, see Geraghty 1983,1986. An additional source is Hooper (in press).
21   It should be noted that there are exceptions. I know of two instances of Tongan /p/being borrowed as Fijian /b/: TON pulumakau “cattle, beef” — FIJ pulumakau, bulumakau, TON pulelulu “Wednesday”— FIJ (Kadavu, Nadrogā) burelulu; and a number of instances of intervocalic Tongan/l/being borrowed as Fijian/r/: TON pulelulu (just noted), TON polosi “long hair” — FIJ porosi. Another caveat is that it is possible for rules of loan phonemicisation to change over time, as a result of changes in the sound systems of the respective languages, and in the bilingualism of their speakers. For these reasons it is linguistically unlikely, rather than altogether impossible, that Fijian burotu was borrowed from Tongan (or other Polynesian) pulotu.
22   Probably also PSS *j (Levy n.d.), which, I suspect, is a conditioned reflex of *d. I have suggested elsewhere (Geraghty 1983:193) that PSS may have a distinct reflex of PEO *j which is not currently reconstructed, but there is as yet insufficient data from what appears to be the criterion language, Sandfly Passage Nggela.
23   As noted in Geraghty 1990:69, external evidence points to PCP *ije, from PEO *Rije, so prothesis of /s/ or /j/ appears to be a shared innovation of Fijian and Rotuman.
24   The Fijian reflex /b/ of PCP *m is unexpected, but not without parallel, cf. “k fish, Monotaxis grandoculis” < *mū, bubu-c “chew and suck” < *mumu-c.
25   The lengthening of pretonic PCP *a in Fijian is unexpected, but not totally irregular, cf. mācawa “space” < *macawa, vāgo-n “awaken” < *vag(ou)-n, cābutu “k fish, Lethrinus sp.” < *cabutu.
26   Sporadic prothesis of /m/ before back vowels in Rotuman is not unlikely; the prothesis of /r/ and /g/ before initial /a/ is regular (Geraghty 1986:307).
27   The next island west after Matuku is Ono-i-Kadavu, situated north of the eastern end of Kadavu. Although the Kadavu musk-parrot is present (Clunie 1984:64), there is a well-known Fijian legend that explains why the parrot is not found there. It is possible, therefore, that the parrot was also hunted to extinction, or near-extinction, on Ono-i-Kadavu, then re-established itself. Somewhat further afield, there is evidence that Polynesians travelled as far as Vanuatu and Pohnpei in Micronesia in their search for red feathers (Geraghty, in press); perhaps this was in response to their declining availability in Fiji.
28   Papua strikes me as a European-inspired interpolation, since it is not found in an earlier version (Gill n.d.[1876]:27). On the other hand, it does occur as a place-name in parts of Polynesia (Smith 1910:113), and is found in a list of islands visited by the ancestors of the Rarotongans (Smith 1910:171) adjacent to a number of what appear to be Fijian islands. It may refer to Bua, the western end of Vanualevu, which was the sole source of that other commodity dear to the Tongans, sandalwood (Martin 1818,I:309), but the syllable Pa- is unexplained.
29   Contemporary Samoan has fanua, not fenua, so Samoan may not have been the immediate Polynesian source of Kiribati Benuakura; fenua, however, is found in Futuna, East Uvea, Tokelau, and Tuvalu.
30   Vanuakula is also the name of two villages in inland eastern Vitilevu, one in Rä and one in Naitāsiri province, neither of which would appear to be likely homelands for the Polynesians, but may possibly have been sources of red feathers. On the other hand, the island of Vanuakula, the northernmost island of the Kadavu group, could well have been known to the Polynesians. Though small and uninhabited, it is conspicuous, and within the range of both red-feathered birds known to have been traded; moreover, it has legendary associations with Tonga. A well-known legend tells of the Princess of Vanuakula being kidnapped by the Prince of Tonga, and her being rescued by a chief of Kadavu in his magic flying canoe made of kīkī(Flagellaria indica, a cane used for weaving baskets). A version of this legend is recorded as Cokanasiga 1990.
31   If one was settled earlier than the other, a stronger case can be made for Samoa, since Tonga is named from a Samoan perspective, Proto Polynesian *toga meaning “south, south wind” (Biggs 1990). Poulsen (1977:8) also states that both Tongan and Samoan traditions indicate that Tonga was settled from Samoa.
32   Over 50 years ago, Burrows (1938:121) remarked that “Nauru seems to be unique in Micronesia in that it lacks traditions of a former homeland”. From recent intensive comparative studies of Micronesian languages, it does appear that Nauruan is the sole member of a first-order subgroup of Greater Micronesian (Jackson 1986:214), so Nauru may well have been the first home of speakers of Micronesian languages. But it is unwise to assume that traditions of a migration from a former homeland always refer to the initial occupation of an island.
33   The crucial evidence for this reconstruction is: Roviana gurata, Proto North Vanuatu *kura-ti, Fijian kura. There are also reflexes in the South-east Solomons, New Caledonia, and Rotuma.
34   In Polynesian languages, this plant is known by reflexes of Proto Polynesian *nonu, which is also widely reflected in Oceania (e.g., Micronesian, New Guinea), so that it must also be reconstructed for Proto Oceanic. Evidence from mainland New Guinea and the Admiralties (Blust 1978:7, 40,147) indicates that the Proto Oceanic form was *n˜on˜um. Why there should be two names is not at all clear, unless perhaps, as suggested in Polynesia, *n˜on˜um was the name of the plant, and *kurat the name of the dye produced from it.
35   Biggs (1990) lists two meanings: “red” and “bird spp.”. A review of the evidence suggests the following: “red, (skin) light brown; k red-feathered bird; ornamental red feathers”. I am assuming that PPN *kula “circumcision” is a different lexeme, related to Proto Micronesian *kurat “pull back the foreskin, circumcise”, though the colour red is also often associated with circumcision.
36   The earlier Fijian term for the bird Phigys solitarius may have been drisi, which is still used in parts of western Vitilevu and Vanualevu, that is, in each of the two major subgroups of Fijian. PPN *kula, or a reflex thereof, is also loaned in triangle Polynesia's other closest neighbours: Kiribati urababā “bright red colour spreading all over”, urabaka “(ripe fruit) bright red”, urabo “faded red”, uraura “red, vermilion”, all from Samoan 'ula, ‘ula’ ula “red, crimson”, and possibly kura “k small bird”; Rotuman kura “k bird, similar to the tāvāke [tropic bird, Phaethon sp.], but red”, possibly kura “helmet”, and ji 'ura “k ornamental shrub, Cordyline sp., with broad reddish leaves and red leaf-stalks”. This last is also loaned in Fijian as tīkula.
37   The cognates are: Sawaiki, a deserted settlement on the east coast of Vatoa; Nukubolu, an area of eastern Vanualevu on the north shore of Natewa Bay; Olosega, a village on the north coast of Qamea; Taqū, a recently abandoned settlement on the east coast of Kabara (but another cognate, Takū, is the name of a coastal stretch near Makadru on Matuku); and Manuka, one of the highest peaks on Taveuni. The loss of the first syllable in Nukubolu is irregular but not entirely unmotivated; part of the original first syllable appears to be preserved in the Mangaia (Cook Islands) form Ukupolu (Gill 1876:114).
38   I have checked the Rapanui dictionaries of Churchill and Englert, but Fuentes is not available to me here.
39   Fornander (1878:I:34) quotes in a footnote from V. de Roehas (sic, for de Rochas 1859) “La Nouvelle Calédonie”, which is not available to me, that, in ancient times, combined parties of Tongans and Fijians invaded New Caledonia and some established chiefly lines. Fornander goes on to refer to these invaders as “Tonga-Fijians”. It may be that, as happened elsewhere (e.g., Smith 1910:163), the name Tongafiti was simply assumed by the investigator to refer to a mixed group of Tongans and Fijians.
40   There is no such PPN reconstruction in Biggs' (1990) Proto Polynesian file. Elbert (1975:313) and Pukui and Elbert (1986:166) reconstruct PPN *tongafiti, but neither gloss nor evidence has been published. The crucial evidence is: Tongan tongafisi “soothsaying, second-sight, occultism; clever ruse or stratagem”, Samoan togafiti “treat medically; remedy, device, scheme, stratagem, prank, trick”, Tuvalu togafiti “trick”, Rennellese tongahiti “stranger, foreigner, traveller, Melanesian”, Hawaiian konohiki “headman of an ahupua'a land division under the chief; land or fishing rights under control of the konohiki”, Marquesan tokohiti “always recovering from illness; skilled, artistic; chief”, Tuamotu tongohiti “unexpected unknown vessel, hostile war-canoe; enemy, foreigner; meteor”, Mangareva tongo'iti “lord, prince”. This fascinating set of cognates deserves more detailed attention than I can afford it in this paper. Note also that, as with pulotu and fanuakula, there appears to be no Rapanui (Easter Island) reflex; whereas, as with pulotu, there is at least one Solomon Islands Outlier reflex.
41   Usually known locally as Koroitogaviji, meaning Togaviji hill-fort. The palatalisation of/t/ to /j/ ([t∫]) before /i/ is regular.
42   Best's (1984:555) surprising observation that the pig appears archaeologically in Lakeba only about 1,000 years ago receives some support from linguistics, the most widespread Eastern Fijian word for pig, vuaka/puaka, being clearly a Polynesian loan. The Makadru tradition that Burotu submerged because they killed a white pig sent from Burotu, the source of all exotic things, may therefore refer to the introduction of the pig (or, in the case of Matuku, possibly reintroduction), presumably from Tonga, and so constitute further evidence that Burotu declined about 1,000 B.P.
43   Credit must go to Hocart (1929:231-2,1952:4-5) for being the first to apply the principles of loan phonology to the discussion of alleged Polynesian loan-words in Fiji.
44   Blust (1976) believed Fijian /c/ to be a possible reflex of his “third palatal”, in which case Tikopia would not necessarily be a loan from Cikobia. While I accept that there is indeed a third palatal (Proto Central Pacific *j), my view is that the Fijian reflex is /s/ (and probably sometimes /d/), and that the apparent Fijian /c/ reflexes are either spurious, or cases of borrowing.
45   Yanuca is a common name for small islands in Fiji, and has cognates in the South-east Solomons (Arosi anuta, Nggela anuha) and Micronesia (Woleai yaliuta “small uninhabited island”), all ultimately derived from Proto Austronesian *nusa“island”.
46   Perhaps the same applies to the naming of Hawai'i, which somewhat resembles Savai'i in its distance and bearing from Central Eastern Polynesia, relative size, volcanic nature, and general appearance.