- 86

ON the northern side of the island of Upolu, in the Samoan group, about six miles inland from the township of Apia in a southwesterly direction, there are situated a number of lava-tunnels known as the Falemaunga caves. According to tradition these caves were used by the Samoans as a place of refuge during the Tongan invasion; there is still extensive evidence of their prolonged human occupation. This paper is concerned with a description of the caves and a brief discussion of their history.

The island of Upolu, which is centrally situated in the Samoan group, is about forty-seven miles in length and about sixteen miles at its greatest breadth. It lies in a direction from west-north-west to east-south-east, and along this central line—the backbone as it were of the island—is a series of craters and summits, most of them well over 2, 000 feet high. The highest in the range is Fito (3, 607 feet). Issuing from these ancient volcanoes are the great lava-flows which have given Upolu its characteristic topography. It is in one of these old lava-flows—one that has its source near the now extinct volcano Singaele—that the Falemaunga caves are found.

Anderson (“The Volcano of Matavanu in Savai'i”, Journal of the Geological Society, Vol. 66, pp. 621-39, 1910) makes mention of the manner in which lava-tunnels are formed. Referring to the eruption of Matavanu, Savai'i, in the year 1906, he states: “The large, fresh lava-streams soon got crusted over on the surface with solidified lava, and the liquid lava continued to flow underneath. Even at the crater it seldom flowed over the lip, but generally entered holes and tunnels in the sides and flowed underground. The lava-field thus became honeycombed with channels of liquid or pasty lava which occasionally came to the surface and flooded it with fresh sheets of lava”.

Further comment upon the formation of these lava-tunnels is made by Thomson (“Geology of Western Samoa, ” New Zealand Journal of Science and Technology, Vol. 4, pp. 60-61, 1921). He writes: “Such tunnels are a characteristic of basaltic eruptions of the Kilauea type, and have an almost circular cross-section while the flow is active, but when the supply of lava ceases that in the tunnel drains out, leaving only a little in the bottom, so that a cooled tunnel has an almost circular section except for a flat bottom. In any tunnel the height and breadth are remarkably constant, except when branching takes place. Hanging from the roof may often be observed stalactites, which are composed of re-fused lava fluxed by the burning of the inflammable gases discharged from the lava flowing below, and young tunnels also have stalactites on the walls, and coatings of soluable salts, chiefly sulphates. The presence of such hollow spaces would lend themselves admirably to ore-deposition if the rocks enclosing them either contained the more valuable metals in sufficient quantity, or the area came later under the influence of metallogenetic gases or solutions. In discussing this aspect with Dr. Jaggar, of Kilauea, he - 87 surmised that the lava-tunnels were not very permanent features, and were probably filled with lavas from later eruptions, giving rise to intrusive bodies of pipe-like form. In Samoa, however, this does not seem to have been the case, and in the extinct lava-flows, both of Savai'i and Upolu, lava-tunnels are of fairly frequent occurrence. They have, however, been considerably modified by the falling-in of rock from the roof and sides, and have lost their original nearly circular and very regular cross-section”.

Fig. 1.

Illustration

The Falemaunga lava-tunnels conform well to this general description. The entrance to the main system of tunnels lies approximately five and one half miles inland (due south) from Malie, a village on the north coast of Upolu, about six miles to the west of Apia by road. All of the tunnels are situated on the property of Mr. S. V. Mackenzie, of Apia, who maintains at Falemaunga a large plantation for the cultivation of bananas, cocoa, and other crops. The distance from Mr. Mackenzie's house at Falemaunga (which lies at the end of the Tapatapao road) to the caves is about three hundred yards in a westerly direction. In September, 1941, I was fortunate in being able to make use of Mr. Mackenzie's house for a period of two weeks, during which time a thorough investigation of all of the caves was made. A preliminary visit of three days was undertaken in July, 1941, for the purpose of reconnaissance, and a brief final inspection of the area in October, 1943.

A survey of both of the main tunnels was first made—a prismatic compass and chain, and an Abney level being used. The entrance to the tunnels is formed by a downbreak in their course. This down-break is almost circular, and measures approximately 50 feet in diameter. The initial station of the compass traverse was established on the eastern lip of this downbreak (Fig. 2). The height above sea-level at this point (as measured with an aneroid barometer) averaged 1, 555 feet. The downbreak has divided what was once a single tunnel into two distinct sections—one to the north and one to the south. Of the two, the northern section is by far the longer. This northern section I have called the north cave, while the shorter southern section I have called the south cave.

- 88

The entrance to the north cave lies 25 feet below the edge of the downbreak, and is extremely confined being only four and one half feet in height and three feet in width. The cave could thus be entered by only one person at a time, and would prove quite impregnable to direct attack. The distance from the entrance of the north cave to its termination (stations 1-19) proved to be 1, 408 feet; the termination being formed by a large downfall of lava which seems to be of relatively recent origin. The width of the tunnel is at first very regular and averages at 25 feet, but at above 850 feet from the entrance (near station 12) there is an amphitheatre of considerable size which forms the centre of a series of branches. The first of these branches is at station 20 (the magnetic bearing being 201 degrees), this branch is very confined, however, and proceeds for a distance of only 150 feet. The second branch, which proved to be a major one, commences at station 12. Its course beyond station 23, where all signs of human habitation cease, is marked in the map by a dotted line. Its direction is at first toward the south-west but at a distance of about 100 feet from station 23, the tunnel takes a sharp turn to the north. From here on its course is very uneven and rough. This branch was followed about 300 yards beyond station 23 at which point the danger of falling rock caused the exploration to be discontinued. The tunnel was extremely confined at this juncture, and there is no definite evidence concerning the distance it extends beyond this furthest point reached.

The maxium width of the main north tunnel at the amphitheatre mentioned above is 56 feet. Just beyond here there is a deviation in the course of the tunnel; the western deviation (stations 13 b. and 14 b.) being at a slightly higher level than the main tunnel (stations 13 a. and 14 a.). The average width of the remainder of the main tunnel (stations 15-19) is 31 feet. The height of the north cave varies from 4 feet at the entrance to over 30 feet near station 12.

The entrance to the south cave, in contrast to that of the north, is over 30 feet wide and about 20 feet high. The distance from the entrance of the south cave to its end is 513 feet, and there is no branching.

The downbreak by which the tunnels are approached still stands in dense bush. According to Thomson (“The Geology of Western Samoa, ” New Zealand Journal of Science and Technology, Vol. 4, p. 61, 1921), who visited the Falemaunga caves in 1920, “the rock is a scoriaceous basalt with the typical pahoehoe type of vesiculation”. Neither of the tunnels is in its primitive state, for there has been much falling in of the roof and sides. Nowhere do the roof and sides seem to be the original ones, and lava stalactites are absent. There are, however, small stalactites one inch to two inches long of a soft white earthy material evidently arising from the weathering of the superincumbent rock. Both of the caves are now inhabited by numerous pe'ape'a (Collocalia francica), a species of swift; and by many small bats. In some places, especially in the south cave, there are deposits of guano over a foot thick.

The caves still contain abundant evidence of human occupation. In both of them are to be found elaborate systems of platforms somewhat similar to those reported as existing in the Seuao caves. 1

In all, 152 platforms were recorded, 129 in the north cave, and 23 in the south cave. These platforms have all been marked on the map. - 89 Unlike the terraces in the Seuao cave, which were all on the same level as the floor of the cave, the platforms in the Falemaunga caves are all built up to a height of from two to three feet. This has been necessitated by the uneven nature of the floor of the tunnels. The platforms are constructed in the main from pieces of fallen lava-rock. In most cases only two retaining-walls have been built, for wherever possible advantage has been taken of the natural contours of the tunnels. In the centre of each tunnel there is a carefully formed pathway.

One of the better constructed of the platforms — the one immediately north of station 13 b. in the north cave, has been selected for more detailed discussion. It may be considered as typical. This platform (No. 127 in my field notes) is built against the eastern wall of the cave, and two retaining-walls only have been used—one at the side and one at the end. Its dimensions are 10 feet in length by 9 feet in width, and its height above the floor of the cave varies from 2 feet at one end to 3 feet at the other (see Fig. 3). Much care has been taken in the construction of this platform. The pieces of lava-rock composing the walls have been neatly and securely fitted together and upon excavation these walls proved to be over 2 feet in thickness. The surface of the platform consists of fine lava rubble, and closely resembles the floor of an ordinary Samoan fale. The surface of some of the other platforms, however, merely consists of a rough mosaic of large pieces of lava.

The platforms commence immediately inside the north cave and continue without intermission to the point at which the tunnel is blocked by a downfall. Much of the debris from this downfall extends over the last visible of the platforms. It would seem certain that the tunnel and the platforms once extended beyond this point. A search was made of the surface of the lava-flow (the position being determined by reference to the data obtained from the survey), to ascertain if there were any other entrances or exits to the tunnels—especially beyond station 19, but none was found. The platforms in the south cave extend from just inside to a point about 85 feet from the end, at which point the cave becomes too narrow and steep to permit of any more.

The Falemaunga plantation was formerly the property of a German, Herr Paul Ludwig Schroeder. His son, Mr. H. H. Schroeder, was still living in Samoa in 1941, and I was able to obtain from him details of the first discovery of the caves. It appears that early in 1914, Mr. H. H. Schroeder was searching in the bush for timber and that he came across the downbreak by mere chance. He subsequently explored the caves and removed from them a number of stone adzes, “some over a foot in length"; and the skull of a child “of almost decayed substance”, which he found on one of the platforms. All of these discoveries were later sent to Germany. The caves were also visited by members of the New Zealand Expeditionary Force which occupied Western Samoa in August, 1914; and by Dr. J. Allan Thomson, Director of the Dominion Muesum, N.Z., in 1920.

Following the completion of the survey a thorough search of both the caves was made for artifacts and other remains; and a number of excavations were undertaken. For this purpose a party of three Samoans was employed as labour during the whole period. Conditions of work were not easy, for the darkness was complete and the humidity extremely high. Two 300 candle power kerosene lamps and several electric torches were constantly used.

- 90

A great number of umu-sites, fire-places, and kitchen-middens were found. In some of the fire-places the ash deposit was well over a foot thick. Excavation of these sites resulted chiefly in the discovery of pig-bones and teeth 2 3 and shell-remains in considerable quantities.

The pig-bones and teeth were almost without exception of immature animals. The following varieties of shell-fish were identified:

All of these shell-fish are edible and all are found in the lagoons and on the reefs of the north coast of Upolu. It is worthy of note that the lagoons of Faleata and Saleimoa, those nearest to the caves, are between five and six miles away and that the shell-fish, which were certainly used as food, must have been carried for at least this distance.

Several lumps of 'ele were also found. 'Ele is a type of red volcanic earth and is generally valued by the Samoans. It is used for the purpose of marking siapo, or tapa cloth. On the island of Upolu,

- i

Plate 1.

View of platform north of station 136, north cave.

Plate 2.

Another view of platform in north cave.

- ii Page is blank

- 91

'ele of good quality is obtained from one place only, an isolated bay on the north coast called Uafato. From here the 'ele still passes by way of trade to the districts which have none. This was apparently the custom also in ancient times, for it must have been in this manner that the lumps of 'ele found their way into the Falemaunga caves.

Five stone adzes, and a number of adze fragments, were also discovered in the caves. They were all found buried either in the shallow surface rubble on the platforms (Figs. 4, 5, and 6); or in kitchen-middens in association with shells and pig-bones (Figs. 7 and 8). One was found in the south cave (Fig. 4), while the other four are all from the north cave. All of these adzes are small in size, the larger specimens having been removed by the Schroeder family and others. Two of these larger adzes were obtained by Mr. Roger Duff, Ethnologist, Canterbury Museum, during his stay in Samoa, from Mr. Swalger, a half-caste German planter of Falemaunga, who stated that they had been found in the caves. I have figured these two adzes in addition to the five mentioned above (Figs. 10 and 11).

All of the adzes, according to the typology of Samoan adzes devised by Te Rangi Hiroa (Samoan Material Culture, Bernice P. Bishop Museum, Bulletin 75, Hawaii, 1930, p. 332 f.) are of Type 1. This is by far the most common class of adze found in Samoa, and may be fairly considered as the Samoan proto-type. The distinguishing features of the type have been well laid out by Te Rangi Hiroa (Ibidem., pp. 334-5): “The type is quadrangular and narrows markedly from edge to poll. The front is not so wide as the back and narrows more progressively towards the poll so that the angles formed with the sides become more obtuse as the poll is approached. The surface is well ground but may be rough towards the poll. The sides are fairly evenly chipped and usually not ground which results in the front edges being irregular. They are, however, ground towards the lower ends of the cutting edge, while sharp edges left by chipping may be removed by grinding. The bevel is quadrilateral but the chin defining its upper border may be irregular owing to the unground nature of the back, or it may be ground off. No distinctive line occurs between blade and butt though the butt part is usually less carefully chipped. The poll may be trimmed to form a surface which slopes upwards and backwards. The front is convex longitudinally and usually slightly so transversely. The actual front surface, owing to coarse chipping at the sides of the butt, may be narrowed to a point, thus making the surface triangular in shape. The chipping of the back is fairly level but it is convex longitudinally and slightly so transversely. The edge is straight but may be slightly raised at the ends. In most adzes the edge is the widest part of the adze, but in a few it is slightly narrowed by grinding the lower end of the sides.” There is considerable variation of size within the type—the specimens recorded in this paper, for example, ranging from 60 mm. to 195 mm. in length.

Te Rangi Hiroa (Ibidem., p. 356) is of the opinion that this common form of Samoan quadrilateral adze (Type 1) with the sides converging towards the front is a primitive type as compared with the dominant form of quadrilateral adze in other areas of Polynesia, in which type a broad cutting-edge is obtained by using a reversed quadrangle with the sides converging towards the back. He claims that “for obtaining a broad cutting edge, both the Hawaiian and New Zealand quadrangles are improvements on the Samoan quadrangle”; and suggests that the fact of the occurence of the reversed quadrangle - 92 in his Type 5 of Samoan adzes (a type very much less common than Type 1) is evidence that the Samoans were on the way to obtaining the broad cutting-edge of other areas.

Support for Te Rangi Hiroa's view of Type 1 as being a primitive type is given by the fact that all of the adzes found in the Falemaunga caves were of Type 1. Tradition affirms that the Falemaunga caves were a place of refuge during the Tongan invasion of Samoa, and if it is assumed that the adzes excavated there date from that period, it is possible to determine their approximate position on a time scale, for the final expulsion of the Tongans from Samoa by the two brothers Tuna and Fata is estimated as having probably occurred during the thirteenth centuary A.D. I have been able to discover no evidence in native tradition of the Falemaunga caves having been used at a period subsequent to the Tongan occupation.

A further object unearthed in the north cave is worthy of mention. It is a small circular‘rubbing’-stone (Fig. 9), which was found lying in the surface-rubble of one of the platforms. It was probably used for polishing or sharpening.

Examination of the ash-deposit found in the numerous fire-places and umu-sites resulted in the ash of the following plants being identified:

There was, however, much other ash which, because of its condition, could not be identified. The remains of coconut-shells (ipu) and husks (pulu) were present in most of the deposits. The wood of the fu'afu'a is commonly used as the lower portion (si'anga) of the Samoan ‘fire-plough’.

The identification of the various types of ash mentioned above, enables an opinion to be formed concerning the kind of lighting used in the caves. The majority of the platforms are well removed from the little light which is admitted by the entrances to the caves. Beyond station 3 in the north cave, and station 4 in the south cave, there is in fact a complete absence of light of any kind. It follows therefore that all of the platforms beyond these two points must have been constructed with the help of some form of artificial illumination, which furthermore would be quite indispensable to anyone living in the caves.

Fires would probably have been the most common form of lighting. Coconut-shells (ipu), which were found in considerable quantities, offer an excellent light when burnt, and are still widely used in Samoa for this purpose. In addition to the light of a fire several other forms were probably used. The leaves of the niu vao (Clinostigma, Wend-land) and maniuniu (Solfia samoensis, Rechinger), both mentioned above, make good 'aulama (torches). Both are readily obtainable in the bush. The dry fronds of the niu (Cocos nucifera, Linnaeus) are much preferred to those of either the niu vao or the maniuniu, but the coconut does not grow at such a height above and distance from the sea, and its fronds would have had to have been transported for a distance of about five miles (as would coconut-shells too; for these, however, there is no good substitute). Because of this the niu vao and maniuniu were probably used instead of the niu. These would not, however, form a permanent source of light; they would be used only as torches while moving about inside the caves. Two other kinds of - 93 lighting were used by the Samoans in pre-European times—the molī popo and the molī lama. The molī popo consisted of a receptacle formed by the half shell of a mature coconut (with the husk removed but the flesh left in to protect the shell), filled with coconut oil (suā u'u) and having as a wick a piece of undyed bark cloth (lau u'a) wrapped around a dry coconut leaflet midrib (tuāniu). The molī lama consisted of a number of the nuts of the candlenut-tree or lama (Aleurites moluccana, Linnaeus) threaded on a dry coconut leaflet midrib. The nuts are first cooked in an oven (umu), and cracked to extract the kernels. Both of these types of lighting are quite efficient, though the light they give is far from strong.

Although no account of cave-burial in Samoa has, to my knowledge, been recorded, it was necessary to bear this possibility in mind during the investigation, for it is well established that cave burials did occur in both Tonga (McKern, Archaeology of Tonga, Bernice P. Bishop Museum, Hawaii, 1929, pp. 90-2) and Fiji (Williams, Fiji and the Fijians, London, 1858, vol. 1, p. 186). Eight of the platforms and several mounds of lava-rubble each in a separate area of the north cave were accordingly excavated. Most of these excavations proceeded to a depth of above four feet before the floor of the tunnel was reached. In all cases it was found that the platforms were constructed of pieces of lava—in some cases quite massive, overlaid with lava rubble. In no case was there any evidence of their having been used for the purpose of burial.

These excavations did, however, serve to demonstrate the immense amount of work which must have gone into the construction of the platforms. When all the circumstances are taken into consideration one cannot fail to be impressed by the fact that such a great number of laboriously-constructed platforms should be found at such a distance underground, for some of them lie over one-quarter of a mile from the entrance to the tunnel. The platforms were, it seems certain, used by those who sought refuge in the caves, as sleeping- and living-places. Some of the platforms are more roughly formed than others, and they have been used as umu-sites and fire-places. Others have had more care exercised in their construction and are usually covered with lava rubble about the size of small pebbles. These were probably used as sleeping-places. It is noticeable that most attention has been devoted to the platforms around about the amphitheatre in the north cave (station 12). This area, spacious as it is, may be considered as corresponding to the malae of a village, and it is probable that the platforms there were reserved for chiefs and their families. Living-conditions in the cave must have been difficult. The darkness was complete, and the humidity high (Appendix 2). Food had to be transported over considerable distances. Firing and lighting must have been constant problems. Water, however, was easily obtainable, for there is a small stream only a few yards to the west of the main downbreak. All things assessed, the Falemaunga caves are to be considered a distinct achievement in planning and organization. It speaks well for the primitive Samoan that in the face of danger he should show such ingenuity and resourcefulness.

My information concerning the history of the Falemaunga caves was derived in the main from the orators (tulafale) and chiefs (ali'i) of Faleata, Tuamasanga, my principal informant being the high chief Faumuinā. The present holder of the title Faumuinā is a person of the highest rank, and also holds the two other important titles: that of Mata'afa (from Aleipata) and Fiamē (from Lotofanga). He is - 94 highly venerated by the Samoan people not only because of his rank but because of the part he played as a leader during the time of the Mau. All of my inquiries were carried on in the Samoan language.

According to my informants, the Falemaunga caves were used as a place of refuge by the people of Tuamasanga (the district occupying the central portion of the island of Upolu, and of which Faleata is a part) during the invasion of Samoa by the Tongans. Much detail has naturally been forgotten, for the Tongan invasion took place over 700 years ago, 4 but Faumuinā related that the caves were used as a place of hiding by the women and children and the old men, while a fort was built by the warriors some little distance away to offer resistance to the Tongans should they attempt an offensive. He described the approximate position of the fort, and said that it was called Taua-le-o'o (which literally means “the war which did not eventuate”) for the reason that the Tongans never actually made an attack, and the people of Faleata were never subjugated.

During a subsequent search I was successful in locating the remains of the old fort. The remains are situated at a place called Le'aupuni, and lie about one-half of a mile to the east of the Falemaunga caves. The main feature of the remains is a massive embankment, standing at the head of the wide valley and just above a confluence in the course of the Le'aupuni stream, which is a tributary of Fuluasonu river. The embankment is composed entirely of red volcanic earth. It measured about 100 feet in length, and over 20 feet in height and had a very steep face. On the ridges behind the embankment both to the east and the west further remains of fortifications were discovered, chiefly in the form of trenches.

My sincere thanks are due to Mr. S. V. Mackenzie, of Apia, Western Samoa, for his kindness in permitting me the use of his house at Falemaunga during the course of my investigations; to Mr. J. Radford, and Pilot Officer J. D. Coulter for help in making the survey of the caves; to Mr. R. Duff for permission to figure two adzes collected by him in Samoa; and to Dr. H. D. Skinner for both advice and assistance.

I am indebted to Pilot Officer J. D. Coulter for the photographs.

- 95

• 1. View of platform lying immediately to the north of station 13 b. in the north cave. This is the platform shown in Fig. 3. The photograph is taken from below and facing the eastern wall of the cave.
• 2. Another view of the same platform taken from above and looking north.

Both photographs are time-exposures taken with the aid of a 300 c.p. kerosene pressure-lamp.

• 1. Map of the island of Upolu, Samoa, showing the location of the Falemaunga caves.
• 2. Map of the Falemaunga caves, showing platforms.
• 3. Plan, and side and end elevations of a typical platform. The platform figured lies immediately north of station 13 b. in the North Cave.
• 4. Adze. Otago Museum, N.Z. D.44.45.
• 5. Adze. Otago Museum, N.Z. D.44.46.
• 6. Adze. Otago Museum, N.Z. D.44.47.
• 7. Adze. Otago Museum, N.Z. D.44.48.
• 8. Adze. Otago Museum, N.Z., D.44.49.
• 9. Stone. Otago Museum, N.Z. D.44.50.
• 10. Adze. Canterbury Museum, N.Z. E.138.356.
• 11. Adze. Canterbury Museum, N.Z. E.138.357.

• 1. Anderson, T., “The Volcano of Matavanu in Savai'i,” Quarterly Journal of the Geological Society, vol. 66, 1910, pp. 621-39.
• 2. Armstrong, John S., Handlist to the Birds of Samoa, London, 1932.
• 3. Behrens, Reise durch die Sudlander und um die Welt, Frankfurt and Leipzig, 1737.
• 4. Bougainville, Louis Antoine de, Description d'un voyage autour du monde par la fregate “La Boudeuse,” 3 vols., Paris, 1771.
• 5. Christophersen, Erling, “Flowering Plants of Samoa,” Bernice P. Bishop Museum, Bulletin 75, Hawaii, 1930.
• 6. Ella, Rev. S., “The War of Tonga and Samoa and the Origin of the title Malietoa,” Journal of the Polynesian Society, vol. 8, pp. 231-4, New Plymouth, 1899.
• 7. Freeman, J. D., “The Seuao Cave,” J.P.S., vol. 52, pp. 101-9, Wellington, 1943.
• 8. Gifford, E. W., Tongan Society, B.P.B. Mus. Bul. 61, Hawaii, 1929.
• 9. Handbook of Western Samoa, Wellington, N.Z., 1925.
• 10. Kraemer, Augustin, Die Samoa-Inseln, Stuttgart, 2 vols., 1902.
• 11. Lesson, P.A., “Traditions des Iles Samoa,” Revue d'Anthropologie, vol. 5, 1876.

- 96

• 12. McKern, W. C., Archaeology of Tonga, B.P.B. Mus. Bul. 60, Hawaii, 1929.
• 13. Milet de Mureau, Voyage de La Perouse autour du monde, 4 vols., Paris, 1797.
• 14. Pratt, George, Grammar and Dictionary of the Samoan Language, 4th ed., Samoa, 1911.
• 15. Skinner, H. D., “Maori Adzes from the Murihiku Region, New Zealand,” Proceedings of the Third Congress of Prehistorians of the Far East, Singapore, 1938.
• 16. Smith, Percy, “Hawaiki, the whence of the Maori,” J.P.S., vol. 8, p. 6, 1899.
• 17. Stuebel, O., Samoanische Texte, Berlin, 1896.
• 18. Te Rangi Hiroa, “Samoan Material Culture,” B.P.B. Mus Bul. 75, Hawaii, 1930.
• 19. Thomson, J. Allan, “The Geology of Western Samoa,” New Zealand Journal of Science and Technology, vol. 4, pp. 49-66, 1921.
• 20. Tregear, Edward, The Maori-Polynesian Comparative Dictionary, Wellington, N.Z., 1891.
• 21. Turner, George, Samoa a Hundred Years Ago and Long Before, London, 1884.
• 22. West, Thomas, Ten Years in the South Central Polynesia, London, 1865.
• 23. Williams, Thomas, Fiji and the Fijians, vol. 1, London, 1865.

(Observation point—North cave, to east of station 12, approximately 850 feet from entrance.)

- 97

Fig. 2. Falemaunga caves.

Illustration

- 98

Fig. 3. Sketch of typical platform.

Illustration

- 99

Fig. 4.

Illustration

- 100

Fig. 5.

Illustration

- 101

Fig. 6.

Illustration

- 102

Fig. 7.

Illustration

- 103

Fig. 8.

Illustration

- 104

Fig. 9.

Illustration

- 105

Fig. 10.

Illustration

- 106

Fig. 11.

Illustration