- 323

Myths, legends and oral tradition are an integral part of many cultures. In particular, many Melanesian and Polynesian societies have developed vast repertoires of tales, myths and legends since their beginnings. Bascom (1965:26) emphasised that the study of oral tradition gives important insights to the make-up of the cultures, an idea which unfortunately has been largely neglected in many ethnographic studies.

During the past few decades several attempts have been made to interpret myths and legends in terms of natural phenomena. On the basis of widespread traditions of “times of darkness” in the Central Highlands of Papua New Guinea, Brookfield (1961), Watson (1963), Glasse (1963), Blong (1975) and Allen and Wood (1980) suggested that the stories describe a major volcanic eruption and the phenomena that accompanied it. More recently, Blong (1982), using both oral tradition and scientific evidence, concluded that the legends originated at the time of the occurrence of a major volcanic eruption many generations ago. Furthermore, he identified the extent of the effects, the location of the volcanic centre and suggested a probable date for the eruption. Another myth from the Madang area of Papua New Guinea (Mennis 1981) suggests that a volcanic island, Yomba Is., may have existed off the north coast some eight to ten generations ago but is now below sea level.

One such Polynesian society, the Kingdom of Tonga, has based many of its cultural beliefs on the wealth of oral tradition. Gifford (1924) and Collocott (1928) drew together many of the myths, legends, tales and poems of the Tongan people that were undocumented until early this century. These studies, together with several other contributions, e.g. Mahony (1915), Brown (1916), Collocott (1921, 1924), have formed the basis for historical studies.

Since many myths and legends 2 may be based on real events, they may be considered important sources of factual information and thus, valid conclusions can sometimes be drawn from their content. Latukefu (1968:143) stated that, following careful evaluation, Tongan oral tradition has made the history of Tonga:

- 324

Furthermore, Rutherford (1977:27) has stated that:

In the light of these statements, this paper examines a legend common to several of the outlying northern islands of the Kingdom of Tonga, and, drawing on volcanological and meteorological evidence, suggests a probable reason for its origin.

The islands of Niuafo'ou, Niuatoputapu and Tafahi are the northern-most outliers of the Kingdom of Tonga (Fig. l). Niuafo'ou, at latitude 15°36′S; longitude 175°3′ W, lies some 650 kilometres north of Tongatapu, the principal island of Tonga. Niuatoputapu and Tafahi lie 175 kilometres to the southeast of Niuafo'ou at latitude 15°57′S, longitude 173°4′ W.

Niuafo'ou is the almost circular summit of a large submarine volcano (Fig.2A). Covering an area of 57 km2, it consists of a series of lava flows which are, in part, interbedded with pyroclastic materials. 3 The centre of the island is occupied by a caldera containing two lakes, Vai Lahi and Vai Si'i. The geological development of the island has been dominated by volcanic activity, with at least 11 eruptions occurring since 1800 (Taylor 1986a: 123; 1991:8). The most recent activity occurred during March 1985 (SEAN 1985:6-7), but the eruption which had the most significant and lasting effect on the native population occurred in September 1946. This catastrophic event destroyed much of the administrative village of Angaha (Fig.2A) on the north coast and resulted in the total evacuation of the island (Rogers 1986). Following repeated requests from many of the Niuafo'ouans, the Tongan Government finally allowed permanent resettlement to occur in 1958 (Rogers 1981:162).

Niuatoputapu (Fig.2B), although having shown no signs of volcanic activity during historic time, is of volcanic origin. Covering an area of only 15 km2, it consists of an eroded volcanic spine surrounded by a Pleistocene terrace of volcanic detritus and an apron of uplifted marine sediments. Tafahi (Fig.2B: insert), 7 km to the north of Niuatoputapu, is a recent volcanic cone. Although it is densely vegetated, it is probably much younger than Niuatoputapu. Cunningham and Anscombe (1985:228) have suggested that Tafahi may represent a late stage parasitic cone formed on the periphery of the remnants of the ancestral Niuatoputapu volcano.

The islands are politically part of Tonga, but linguistic evidence suggests that Niuafo'ou (Dye 1980:352; Tsukamoto 1988:6-16) and Niuatoputapu (Kirch 1984:233) may have once belonged to the Samoic-Outlier subgroup,

- 325

Figure 1. Map of the Southwest Pacific region showing the location of Niuafo'ou and Niuatoputapu.

- 326

Figure 2.A. Map of Niuafo'ou showing the salient geologic and geographic features and the location of the villages: B. Geomorphic map of Niuatoputapu (after Kirch, 1984) and Tafahi. Outlines and spot heights after Government of Tonga (1975).

- 327

which included islands such as Uvea, 250 km to the north-northwest of Niuafo'ou (Fig. 1). This influence may have lasted until, at least, the time of the first European contacts in 1616 AD. Following detailed archaeological excavations, Rogers (1974:339) and Kirch (1978:12-13) concluded that Niuatoputapu was probably inhabited by Lapita colonisers by early in the first millenium B.C. Although no archaeological surveys have been conducted on Niuafo'ou, it could be expected that from its location it may have been initially colonised during a similar time period to that of neighbouring Niuatoputapu.

The legends that will be considered in this paper relate to the exploits of Sekatoa who was described by Mahony (1915:116) as the:

The first version of the legend was given by Mahony (1915:117):

Rogers (1975:417) has shown that other versions of the legend recorded by him during 1969-70, vary only with respect to the part of the anatomy which Sekatoa used to initiate the impromptu sunrise.

Another version of the same legend was recorded on Niuafo'ou in 1983 (Tsukamoto personal communication 1985):

“There was a Samoan teevolo named Moso. He came from Samoa and took a fancy to the Niuafo'ouan teevolo 'Alaki Vai Lahi's dwelling place and asked for it so that he could take it back to Samoa and live in it. Then 'Alaki Vai Lahi asked Moso what he was supposed to live in if he took his dwelling place away. The Samoan teevolo answered that if he took it away he and 'Alaki Vai Lahi would be able to make another dwelling place for 'Alaki Vai Lahi on Niua. The 'Alaki Vai Lahi told him that he could take it and then - 328 come back to make him a dwelling place. The Samoan teevolo Moso went to fetch other Samoan teevolos. Before long, the Samoan teevolos came to dig up the dwelling place, Tafahi. When they were digging it up-when it got loose from the ground, a big piece of earth fell from it into Vai Lahi, and Moso planted on it a faa kula tree which he brought from Samoa so that 'Alaki Vai Lahi could live in it. Then the teevolos dragged the dwelling place all the way to the seashore. This place is called Fakaholonga Tafahi. The teevolos then swam with it as far as Niuatoputapu, and when they went past, a teevolo named Sekatoa of Niuatoputapu noticed it. Sekatoa went to gather Niuatoputapuan teevolos and ordered them to crow out like roosters. Then he went to the horizon and stared from where he was. The Samoan teevolos thought that the sun had risen and Moso's staring eyes also became bright. They threw Tafahi away in Niuatoputapu and fled to Samoa. So 'Alaki Vai Lahi's dwelling place could not be carried to Samoa and was left in Niuatoputapu. 'Alaki Vai Lahi got his dwelling place, the faa kula tree. Since then, the faa kula tree has been there and we go to Vai to fetch faa kula fruit. A faa tea tree was also planted but it does not bear fruit on this island and just stands there to this day. On a calm day, when there is no breeeze, if we go and stand on a path on the edge of a cliff, we see something twirling in Vai Lahi like dust in a whirlwind on the mountain. Then we know that 'Alaki Vai Lahi is walking on Vai. This is still part of our living tradition and people still believe it. The big faa tree is still there because Moso hasn't got it back. At midnight, when people hear a young, green coconut falling to the ground with a thud, they think that it is Moso, and people call out to him, telling him to bring the take of the coconut for them. Then Moso gets angry because they are asking for the best part of his coconut and goes to the eastern part of the island to eat the coconut there. Ever since, Tafahi has been in Niuatoputapu and 'Alaki Vai Lahi has been living in the faa kula tree. But Moso doesn't have anything.”

The legends presented here, although differing in minor detail, refer to the removal of the central part of Niuafo'ou island and its subsequent relocation near Niuatoputapu and a period of darkness lasting for an unknown period of time that was ended by an impromptu sunrise.

Legends with this theme are not unique to Niuafo'ou and Niuatoputapu. Gifford (1924:89) documented a tale that referred to the origin of Kao, a recent volcanic cone several kilometres north of the active volcanic island of Tofua (Fig. l):

“Three deities from Samoa, Tuvuvata, Sisi and Faingaa, conspired to steal Tofua. So they came and tore up the high mountain by its very roots and its place was taken by a large lake. This enraged the Tongan gods very much and one of them, Tafakula, essayed to stop the thieves. He stood on the island of Luahako and bent over so as to show his anus. It shone so brilliantly that the Samoan deities were struck with fear, thinking that the sun was rising and - 329 that their dastardly work was about to be revealed. Hence, they dropped the mountain and fled to Samoa. The mountain became the island of Kao.”

Gifford (1924:90) also documented another version which varied little from the initial tale:

To a volcanologist, certain aspects of the legends presented here suggest a similarity to phenomena that accompany volcanic eruptions.

Since this investigation involves an analysis of data using an interdisciplinary approach, it is appropriate to outline and discuss several important topics, e.g. the relation between myth and reality together with pertinent volcanological and meteorological aspects, which will have a significant bearing on the conclusions drawn.

Although it is generally accepted that myths, legends and oral traditions may be based on factual events, the accuracy cannot always be easily determined. In the case of Tonga, no strict measures have been adopted to preserve the accuracy of the traditions and hence they are prone to variation. Latukefu (1968:139-40) suggested that they are still of great value to the historian provided they are used cautiously and time is taken to detect sources of distortion. A volcanologist familiar with the interpretation of oral traditions may be able to suggest that certain aspects of the traditions relate to “factual” geohistorical events, e.g. a volcanic eruption that occurred, generations ago, in the northern Tongan region.

The mechanisms and products of volcanic eruptions have been described elsewhere (Macdonald 1972; Bullard 1976, 1979; Williams and McBirney 1979; Fisher and Schmincke 1984; Heiken and Wohletz 1985; Cas and Wright 1987), and so only aspects considered relevant to this discussion will be considered here.

When a volcano erupts as a consequence of excessive pressure within the magma chamber beneath, it produces a mixture of gas, pyroclastic material (tephra) and lava. Depending on the type and chemical composition of the volcano, the proportions of the three components will vary considerably

- 330

- 331

(Table 1). During an explosive eruption mainly gas and tephra are formed. The exsolution of magmatic gas or the formation of water vapour by the instantaneous conversion when water comes in contact with rising magma forms the basic mechanism for the initiation of explosive eruptions. The tephra formed as a result of the eruption is entrained in a developing eruption column which initially rises vertically above the vent.

During the dispersal phase and as the density of the plume declines, the larger and heavier pyroclasts 4 fall from the expanding cloud. The remaining low density suspension of fine tephra disperses laterally under the effect of the prevailing winds, sometimes covering extensive areas around the volcano. The large magnitude eruption of Krakatau, during 1883, affected an area to the west of the volcano estimated to be 827,000 km2 with an ash layer of between 1-5 mm in thickness (Verbeek 1885) 5. Similar cases are also known following smaller magnitude explosive eruptions of basaltic volcanoes. The eruption of Fernandina in the Galapagos Islands during 1968 (Simkin and Howard 1970:431) produced a tephra cloud that was reported 350 km west of the volcano. The 1886 eruption of Niuafo'ou produced a tephra cloud that may have affected an estimated area of 670 km2 to the west of the island (Taylor 1991:39). The distribution of tephra in all these cases was dominated by the prevailing winds. Roobol et al. (1985:325-31) have suggested that the prevailing winds have played a major role in the distribution of tephra in the Lesser Antillies.

Blong (1984:32) has shown that following the production of a tephra cloud, periods of extreme darkness may occur. He cites the 1815 eruption of Tambora, where an eyewitness (Anon. 1816:252-53) described the darkness in the following manner:

Simkin and Fiske (1983:91-105) also noted that a series of pronounced atmospheric phenomena followed the 1883 eruption of Krakatau and that periods of extreme darkness also occurred during the period of the eruption. Although both the Krakatau and Tambora examples relate to eruptions of volcanoes with intermediate to silicic compositions, these events serve to indicate that pronounced periods of darkness may occur during large-scale volcanic eruptions. Simkin and Howard (1970:431) describe the instance following the Fernandina eruption when at a distance of greater than 135 km visibility was down to 500 metres, thus implying a period of darkness. Furthermore, Jaggar (1935:95) noted that during the 1886 explosive eruption of Niuafo'ou the clouds of dust shut off the light of day on many occasions during the 18 day eruption.

- 332

Figure 3. Plots of monthly average upper-air wind velocities and directions for Nandi, Fiji. Data from de Lisle (1969).

- 333

Niuafo'ou and Niuatoputapu are in a region of the south-west Pacific affected by the trade winds. The average surface winds for this region blow from the east and south-east. Upper-air wind profiles for Nandi, Fiji are given in Figure 3.

Figure 3 shows that the troposphere is characterised by zonal winds. Low level winds (1-2 km altitude), blow from the east or southeast while the upper level winds (>3 km) generally blow from the north-west. The upper and lower levels of the troposphere are separated by a trade wind shear zone represented by a zone of low velocity at a height of about 2 km. The dry season winds are more consistent in both direction and velocity than those occurring during the wet season, suggesting that the wet season circulation is affected by the presence of the South Pacific convergence zone (Thompson 1986:6).

The prevailing wind patterns at the time of the 1886 eruption of Niuafo'ou influenced the dispersal of the tephra that was injected into the troposphere (Taylor 1991:48). Thus, during a small scale explosive eruption, ejection of tephra to heights below the trade wind shear zone would be expected, and this tephra would be dispersed along a westerly dispersal axis. However, in the case of a moderate to large scale explosive eruption, ejection of tephra into the upper troposphere (3-19 km in the dry season; and 3-17 km in the wet season) could be expected with the dispersal axis being orientated in an easterly or south-easterly direction.

Computer modelling has been used to predict the probable trajectories of pyroclastic particles released from an eruption column. Using the data given in Figure 3, the model used by Blong (1981:87) has been used to predict the probable dispersal patterns of tephra produced during a moderate to large magnitude explosive eruption that may have occurred on Niuafo'ou in the past. As with the modelling of the 1886 tephra deposit the upper air wind data for Nandi, Fiji (Fig.3), has been used in this analysis. Figure 4 gives examples of the calculated paths of pyroclastic particles released from an eruption cloud at heights of 50, 30, 20, 10 and 5 km during May, August and November. Following release, the particles have been plotted at positions corresponding to 15, 10 and 5 km altitude and at the impact position for each trajectory. Figure 5 shows calculated plots for the trajectories for the wet and dry season averages. These diagrams indicate that, depending on the time of the year during which the eruption occurred, the tephra dispersal and the impact position would be either to the east or to the west. The plot for the dry

- 334

Figure 4. Calculated trajectories for the months of May, August and November for pyroclasts of 0.1mm diameter and a density of 0.5gm/cm3. The trajectories are for particles released at heights of 50, 30, 20, 10 and 5 km above the vent, following the model of Blong (1981). Symbols are described in Figure 5.

- 335

Figures 5a and 5b. Calculated pyroclast trajectories in plan and section for wet and dry season upper-air wind averages.

- 336

Illustration

- 337

Figure 6. Calculated tephra dispersal envelopes for the wet and dry season trajectories given in Figure 5 for release heights of 50, 30 and 20 km only. KK72-7 represents the location of the seafloor sediment core referred to in this paper and described by Taylor (1986b).

- 338

season indicates a predominantly easterly trajectory, while for the wet season, trajectories of greater spread with impact positions to both the east and the west of Niuafo'ou. Tephra dispersal envelopes have been constructed for the average wet and dry season trajectories using the method outlined by Blong (1981:88-92) (Fig.6). The calculated tephra envelopes suggest that particles released from an eruption column produced during moderate to large magnitude explosive activity would disperse in an easterly direction with extensive tephra falls probably occurring in the region of Tafahi and Niuatoputapu.

Vitaliano (1973:104-5) has stated that:

This statement seems to have some credence when the legends concerning Niuafo'ou and Niuatoputapu are considered. Several aspects of the legends suggest that they may have been based on the phenomena occurring during a moderate to large magnitude volcanic eruption. These aspects include:

• i) the removal of the centre of the island of Niuafo'ou;
• ii) the eastward movement of the central portion of Niuafo'ou and its subsequent relocation near Niuatoputapu, as Tafahi; and
• iii) the period of darkness followed by the impromptu sunrise.

The model for the evolution of Niuafo'ou described by Taylor (1991:94-7) involves a caldera-forming 6 phase which resulted in the formation of a large depression at the centre of the island which is now filled by two lakes. This phase involved the removal or subsidence of the central portion of the volcano. It is generally accepted that the most common caldera-forming eruptions result from the subsidence or collapse of a portion of the structure (Williams and McBirney 1979: 207). This phase of activity may explain the aspects of the legends that refer to the removal of the central part of the island.

During caldera-forming eruptions, e.g. Krakatau, 1883 (Simkin and Fiske 1983), or caldera enlargement eruptions, e.g. Fernandina, 1968 (Simkin and Howard 1970), large amounts of volcanic ash are ejected into the atmosphere. Because of the intensity of the eruptions, the eruption column may commonly reach heights in excess of 20 km (e.g. Fernandina 25 km, Krakatau 50 km). Following density stabilisation, the dispersal of the

- 339

- 340

column will commence under the influence of the prevailing winds. In the case of Niuafo'ou, if a large magnitude eruption occurred which produced a tephra cloud that attained a height in excess of 20 km, then using the calculated tephra dispersal patterns given in Figure 6 the prevailing wind patterns would cause a heavy tephra fall to the east of Niuafo'ou in the region of Niuatoputapu and Tafahi. A study of the chemical and morphological characteristics of glass shards present in a sea-floor sediment core recovered from latitude 16°08′S, longitude 175°09.6′W, 60 km to the southeast of Niuafo'ou (Taylor 1986b), has confirmed that tephra with a similar chemical composition to the lavas and pyroclastic deposits that outcrop on Niuafo'ou is present in sea-floor sediments to the southeast of the island (Table 2), thus indicating that a moderate to large scale eruption may have occurred which deposited tephra over an extensive area around Niuafo'ou. This phase of activity may account for the eastward movement of the “central portion” of Niuafo'ou. However, its relocation as Tafahi near Niuatoputapu presents a problem. Geochemically, the rocks of Tafahi contrast with those of Niuafo'ou (Ewart 1976 10; Ewart and Hawkesworth 1987 11), suggesting a different origin for the rocks. Furthermore, the gross geomorphic features of Tafahi, including the size and shape (see Figure 2 for a comparison), also preclude an origin connected with Niuafo'ou. Cunningham and Anscombe (1985:228) have suggested that Tafahi represents a parasitic cone formed during the post-Pliocene development of Niuatoputapu. A possible explanation for the mention of the origin of Tafahi in the legends is that contemporaneous volcanic activity occurred at both sites.

A further aspect of the legend recorded by Tsukamoto (personal communication 1985) and given in an earlier section of this paper:

may also refer to volcanic activity. There are two eruption-related phenomena to which this part of the legend may refer. The first is that it may refer to near-vent tephra fall which would accompany an eruption, but an alternative explanation is that it may refer to the formation of a small cinder or tuff cone within the lake during a later stage of the eruption or during a period of post-caldera explosive activity. As shown in Figure 2, a series of small pyroclastic cones are present within the caldera, and the existence of these features support the alternative explanation.

As noted earlier, Blong (1984:32) and Simkin and Fiske (1983:91-105) have described the occurrence of periods of extreme darkness during explosive eruptions. Similarly, Blong (1982) has described a period of darkness that occurred several generations ago in the New Guinea highlands and which he concluded was the result of the dispersal of the Tibito Tephra, - 341 an ash deposit that was probably produced during a large magnitude explosive eruption of one of the volcanoes along the north coast of Papua New Guinea. Atmospheric phenomena also accompany the dispersal of tephra (see Simkin and Fiske 1983:154-9); those most commonly observed include severe electrical disturbances, and enhanced sunrises and sunsets, or simply the sun being partially or completely obscured by the dispersing tephra cloud. The period of darkness and subsequent sunrise referred to in the legends may thus be explained by a tephra cloud causing partial or complete darkness in the region followed by the reappearance of the sun following dispersal of the tephra cloud.

A further point that must be considered is the timing of the legend and hence the eruption. Since none of the available geological data allows an absolute date to be assigned to the eruption, the use of relative dating techniques must be used. The existence of an as-rich layer in sea-floor sediment as described by Taylor (1986b) may in itself suggest the occurrence of an eruption. Assuming that the presence of this ash, similar in composition to the products exposed on Niuafo'ou, was formed during an eruption that occurred on Niuafo'ou, then using the minimum sedimentation rate of 17 m/Ma 12 given by Brocher et al. (1985:90), an age of late-Pleistocene to early-Holocene may be inferred. However, sedimentation rates may have been extremely variable in the geological environment where the core was recovered, i.e. an active back-arc basin. 13 Churkin and Packham (1973:481) have suggested that during the Pleistocene sedimentation rates may have been as high as 120 m/Ma, averaging 80 m/Ma, which may indicate that the ash layer may have been deposited as recently as 10,000 years ago, however, estimates based on these criteria are far from conclusive.

As noted earlier, legends and oral tradition may contain elements of fact and thus may suggest that the tradition may have been established at the time during which the events occurred. By inference, the event on which a legend has been based must have occurred following the period of colonisation. As noted by Kirch (1978:12), archaeological evidence suggests that this region of the south-west Pacific was inhabited by Lapita colonisers about 3,000 years ago. Thus, the legends and hence the eruption which they describe may have occurred less than 3,000 years ago. However, only detailed studies of the stratigraphic relationships of Lau Basin sediments to the east of Niuafo'ou and on the island of Niuatoputapu will confirm this hypothesis.

Using a volcanological approach, it is suggested that the legends outlined in this paper refer to volcanic activity that has occurred on the island of - 342 Niuafo'ou during a time following the early settlement of the region, possibly less than 3,000 years ago. Volcanological and meteorological evidence and computer modelling suggest that a number of the aspects of the legends can be explained by the occurrence of a moderate to large magnitude eruption which may have resulted in the formation of the caldera but because of inconsistencies in the timing of the events and the origin of the legends, some of the aspects discussed may refer to periods of post-caldera explosive activity.

Another point that is noteworthy is that if the legends referred to in this paper do in fact describe an event that may have occurred many generations ago, it serves to show that the legend has remained in existence for a long period of time and as stated by Latukefu, careful consideration of this form of tradition will indeed make history:

- 343

This paper was prepared as part of a Master of Science project at Macquarie University, Sydney, Australia, hence thanks must go to my supervisor, Associate Professor Trevor Green, for his encouragement throughout the project and for reviewing an early draft of this paper. Thoughtful comments were also provided by Russell Blong, Wally Johnson and an anonymous reviewer, which enabled the manuscript to be improved. Russell Blong and Steve Riley are also thanked for allowing me to use their tephra dispersal computer program. Agahisa Tsukamoto, provided a copy of his thesis on the language of Niuafo'ou and transcripts of several myths/legends recorded during his visits to Niuafo'ou. Thanks must also go to Tevita and Lusia Koloamatangai and their family, who extended there hospitality while I was on Tongatapu. I must also thank Tupou Kata of Kolofo'ou village, Niuafo'ou, for his companionship during my visits to Niuafo'ou. Finally I must thank my “Niuafo'ou family”, Tu'a and Kaufo'ou 'Aholeloi and their children Lisiate, Nasia, 'Ilisa and Matamoana who shared their home, food and Niuafo'ou with me.

• Allen, B. J. and A. W. Wood, 1980. Legendary volcanic eruptions and the Huli, Papua New Guinea. Journal of the Polynesian Society, 89:341-7.
• Anonymous, 1816. Miscellaneous observations on the volcanic eruptions of the islands of Java and Sumbawa, with a particular account of the mud volcano at Grobogar. Quarterly Journal of Science and the Arts, 1:245-58.

• Bascom, W. R., 1965. Folklore and anthropology, in A. Dundes (ed.), The study of folklore, Englewood Cliffs: Prentice-Hall. pp.25-33.
• Blong, R. J., 1975. The Krakatoa myth and the New Guinea Highlands. Journal of the Polynesian Society, 84:213-7.
• —— 1981. Tephra fallout from Karkar volcano: a first approximation, in R. W.

- 344

• Johnson (ed.), Cooke-Ravian Volume of Volcanological Papers, Geological Survey of Papua New Guinea Memior, No. 10, pp.85-93.
• —— 1982. The time of darkness, local legends and volcanic reality in Papua New Guinea. Canberra: Australian National University Press.
• —— 1984. Volcanic hazards: A sourcebook on the effects of eruptions. Sydney: Academic Press.
• Brocher, T. M., S. Wirasantasa, F. Theyer, and C. Mato, 1985. Regional patterns of sedimentation along the northern Melanesian boarderland, in T. M. Brocher (ed.),Geological Investigations of the Northern Melanesian Borderland, Circum-Pacific Council for Energy and Mineral Resources, Houston, Earth Science Series, pp.77-102.
• Brookfield, H. C., 1961. The highland people of New Guinea: a study of distribution and localization. Geographical Journal, 127:436-8.
• Brown, G., 1916. Folktales from the Tonga Islands. Folklore, 17:426-33.
• Brunvand, J. H., 1968. The study of American folklore: an introduction. New York: W.W. Norton.
• Bullard, F. M., 1976. Volcanoes of the earth. St Lucia: University of Queensland Press.
• —— 1979. Volcanoes and their activity, in P. D. Sheets and D.K. Grayson (eds), Volcanic activity and human ecology, New York: Academic Press, pp.9-48.

• Cas, R. A. F. and J. V. Wright, 1987. Volcanic successions modern and ancient: a geological approach to process, products and successions. London: Allen and Unwin.
• Churkin, M. and G. H. Packham, 1973. Volcanic rocks and volcanic constituents in sediments, Leg 21, Deep Sea Drilling Project, in R. E. Burns et al., Initial Reports of the Deep Sea Drilling Project, Washington D.C. Government Printing Office, 21:481-93.
• Collocott, E. E. V., 1921. Legends from Tonga. Folklore, 32:45-58.
• —— 1924. Tongan myths and legends. Folklore, 35:275-83, 372-8.
• —— 1928. Tales and poems of Tonga. Honolulu: Bernice P. Bishop Museum Bulletin, No.46.
• Cunningham, J. K. and K. J. Anscombe, 1985. Geology of 'Eua and other islands of the Kingdom of Tonga, in D. W. Scholl and T. L. Vallier (eds), Geology and offshore resources of the Pacific island arcs-Tonga region. Circum-Pacific Council for Energy and Mineral Resources, Houston, Earth Science Series, pp.221-57.

• de Lisle, J. F., 1969. Upper-air wind statistics for New Zealand stations. New Zealand Meteorological Service Miscellaneous Publication, No. 129.
• Dye, T. S. 1980. The linguistic position of Niuafo'ou. Journal of the Polynesian Society, 89:49-357.

• Ewart, A., 1976. A petrological study of the younger Tongan andesites and dacites and olivine tholeiites of Niuafo'ou Island, S.W. Pacific. Contributions to Mineralogy and Petrology, 58:1-21.
• —— and C. J. Hawkesworth, 1987. The Pleistocene-Recent Tonga-Kermadec arc

- 345

• lavas: Interpretation of new isotopic and rare earth data in terms of a depleted mantle source model. Journal of Petrology, 28:495-530.

• Fisher, R.V. and H.-U. Schmincke, 1984. Pyroclastic rocks. Berlin: Springer-Verlag.

• Gifford, E. W., 1924. Tongan myths and tales. Honolulu: Bernice P. Bishop Museum Bulletin, No.8.
• Glasse, R. M., 1963. Bingi at Tari. Journal of the Polynesian Society, 72:270-1.
• Government of Tonga, 1975. 1:25,000 Topographic map of Niuafo'ou. Series X872, Sheet 1.

• Heiken, G. H. and K. H. Wohletz, 1985. Volcanic ash. Berkeley: University of California Press.

• Jaggar, T. A., 1935. Living on a volcano. The National Geographic Magazine, 68:91-106.

• Kirch, P. V., 1978. The Lapitoid period of west Polynesia; excavations and survey in Niuatoputapu, Tonga. Journal of Field Archeology, 5:1-13.
• —— 1984. The evolution of Polynesian chiefdoms. Cambridge: Cambridge University Press.

• Latukefu, S., 1968. Oral traditions: An appraisal of their value in historical research in Tonga. Journal of Pacific History, 3:135-43.

• Macdonald, G. A., 1972. Volcanoes. Englewood Cliffs: Prentice-Hall.
• Mahony, B. G., 1915. Legends of the Niua Islands. Journal of the Polynesian Society, 24:116-17.
• Mennis, M. R., 1981. Yomba Island: a real or mythical volcano, in R.W. Johnson (ed.), Cooke-Ravian Volume of Volcanological Papers, Geological Survey of Papua New Guinea Memoir, No. 10, pp.95-9.

• Roobol, M. J., A. L. Smith, and J. V. Wright, 1985. Dispersal and characteristics of pyroclastic fall deposits from Mt Misery volcano, West Indies. Geologische Rundschau, 74:321-35.
• Rogers, G., 1974. Archaeological discoveries on Niuatoputapu Island, Tonga. Journal of the Polynesian Society, 83:308-48.
• —— 1975. Kai and kava in Niuatoputapu: social relations, ideologies and contexts in a rural Tongan community. Unpublished PhD thesis, University of Auckland.
• —— 1981. The evacuation of Niuafo'ou: an outlier in the Kingdom of Tonga. Journal of Pacific History, 16:149-63.
• —— 1986. The fire has jumped: Eyewitness accounts of the eruption and evacuation of Niuafo'ou, Tonga. Institute of Pacific Studies, University of the South Pacific, Suva.
• Rutherford, N., 1977. Oral tradition and prehistory, in N. Rutherford (ed.), Friendly Islands: a history of Tonga. Melbourne: Oxford University Press, pp.27-39.

• SEAN Bulletin, 1985. Niuafo'ou volcano. SEAN Bulletin, vol. 10, 9:6-7.
• Simkin, T. and R. S. Fiske, 1983. Krakatau: the volcanic eruption and its effects. Washington D.C.: Smithsonian Institution Press.
• Simkin, T. and K. A. Howard, 1970. Caldera collapse in the Galapagos Islands. Science, 169:429-37.

- 346

• Taylor, B. and G. D. Karner, 1983. On the evolution of marginal basins. Reviews of Geophysics and Space Reviews, 21:1727-41.
• Taylor, P. W., 1986a. Niuafo'ou Island Tonga: subaerial volcanism in an active back-arc basin (abstract), International Volcanological Congress, New Zealand, Abstract volume, p. 123.
• —— 1986b. Compositional and textural variation in volcanic glass from the Lau Basin, SW Pacific: implications for volcanogenic sedimentation in a back-arc basin. Unpublished Postgraduate report, Macquarie University.
• —— 1991. The geology and petrology of Niuafo'ou Island, Tonga: Subaerial volcanism in an active back-arc basin. Unpublished MSc thesis, Macquarie University.
• Thompson, C. S., 1986. The climate and weather of Tonga. New Zealand Meteorological Service Miscellaneous Publication, No. 188(5).
• Tsukamoto, A., 1988. The language of Niuafo'ou Island. Unpublished PhD thesis, Australian National University.

• Verbeek, R. D. M., 1885. Krakatau. Batavia.
• Vitaliano, D. B., 1973. Legends of the Earth: their geologic origins. Bloomington: Indiana University Press.

• Watson, J. B., 1963. Krakatoa's echo. Journal of the Polynesian Society, 72:152-5.
• Williams, H. and A. R. McBirney, 1979. Volcanology. San Francisco: Freeman Cooper.
• Wright, J. V., A. L. Smith and S. Self, 1980. A working terminology of pyroclastic deposits. Journal of Volcanology and Geothermal Research, 8:315-36.