«Exploration and settlement of the far distant islands of the Polynesian triangle are feats of unmatched difficulty within the history of human ...»
TRIANGULAR MEN ON ONE VERY LONG VOYAGE:
THE CONTEXT AND IMPLICATIONS OF A HAWAIIAN-
STYLE PETROGLYPH SITE IN THE POLYNESIAN
KINGDOM OF TONGA
Kanokupolu, Tongatapu, Kingdom of Tonga
DAVID V. BURLEY
Data from a recently discovered petroglyph site at Houmale‘eia on Foa Island in the Ha‘apai Island Group of Tonga are provocative relative to the East/West Polynesian interface and a continued argument for extensive sea-going voyages in later prehistory. The Houmale‘eia motif suite has no developmental precedent in Tonga or West Polynesia. Rather, it is typical of Hawaiian rock art dating to the approximate interval AD 1400 to 1600.
In the following paper we provide an account of this site, its motifs, and its comparative milieu. The site is further contextualised within traditional histories for Tonga and Hawai‘i, and implications for long distance voyaging between West and East Polynesia are examined.
209 210 Triangular Men on One Very Long Voyage Figure 1. Voyaging paths for the settlement of Polynesia based on Kirch (2000: 131). Inset map is of central and southern Tonga. Arrows indicate sites identiﬁed in text.
211 Shane Egan and David V. Burley
THE HOUMALE‘EIA PETROGLYPH SITEThe Houmale‘eia petroglyph site is situated on the northeast bay on the northernmost peninsula of land on Foa Island (Fig. 1). Houmale‘eia is the name given the area by local residents and translates roughly as point of land (houma) pushing or moving to one side (le‘ei) (Churchward 1959: 293), presumably in reference to the shifting sand along its beach. Foa Island is a raised coral limestone formation with a 15-19m maximum elevation occurring on its northern end. Relative sea levels for Foa were up to 1.8m higher than present at c. 900 BC when the first Lapita colonists settled the island (Dickinson, Burley and Shutler 1994). Stabilisation of land/sea relationships and development of the present day beachfront at Houmale‘eia is estimated by c. AD 1. The small bay in which the Houmale‘eia site is situated is protected on the windward side by fringing reef (Fig. 2). A cuspate beach of coral sand covers the near shore reef surface. Beach rock also has formed over the eroded reef-flat in different areas along the shore and it forms the medium on which the petroglyph site was executed. Beach rock is a surface veneer of sand and other beach debris cemented by calcium carbonate precipitate (Crane
1992: 60). Typically it incorporates bedding planes that, in later prehistory, facilitated its removal in blocks for use in tomb construction (Burley 1997).
Three laminae surfaces are present in the beach rock of the Houmale‘eia site.
These tilt downward from shore to water and create a ramp merging with the reef platform (Figs 2 and 3). Houmale‘eia site panels occur within tidal range and are completely submerged at high tide.
Burley (1992) conducted an archaeological survey on Foa Island in 1991, recording several sites associated with the late prehistoric Tongan chiefdom as well as a Polynesian Plainware ceramic period shell midden (c. AD 100on the island’s north end. The petroglyph site was not documented, it likely being covered by coral sand at the time. Completely exposed tree root systems along the present shore inland of the site indicate recent removal and redeposition of the sand blanket by wave action. Exposure of the petroglyph panels is a consequence.
Figure 4. Houmale‘eia western half of upper petroglyph panel, view to south.
Photograph was taken at night with side-light shadowing.
Triangular Men on One Very Long Voyage 214 Figure 5. Houmale‘eia upper (western) panel images.
The petroglyph site has two principal panels or image clusters with additional examples of outlier images (Fig. 3). Site images generally are oriented north/south with most facing the water. Beach stone is softened when wet and execution of surface grooves through abrasion was accomplished without the investment of substantial effort. Softness of the base material has led to surface erosion and many of the image grooves are now shallow, subtle and difficult to see in normal daylight. Further erosion is anticipated through the twice-daily wash of tide and coral sand. To document the site, image groupings were traced onto clear plastic sheets with indelible markers.
Sheet orientations and locations were positioned on a base map. Side lighting with a battery powered torch, and night light photography further allowed image definition and transcription (Fig. 4). Traced images were re-mapped and scaled down in the laboratory both individually and as panel groups (Figs 5, 6 and 7). Overall the motif suite appears minimalistic but figurative and rigidly formal in stylised designs. This stylistic homogeneity speaks of a single or small group of artists working over a limited period of time.
Shane Egan and David V. Burley 215
Figure 6. Houmale‘eia lower (eastern) panel images.
The western upper panel is positioned on upper and intermediate beach rock layers (Figs 2, 3 and 5). Broken images on the inland edge of the beach rock and damage on the periphery of two large cupules indicate that portions of the panel are now destroyed. The panel incorporates upward of 40 images including 22 complete or partial anthropomorphs. The density and clustering of anthropomorphs seemingly document an event in which a group of people was aggregated. A bird motif and what may be a snare suggests the chiefly sport of pigeon snaring, an activity known to have occurred at a pigeon snaring mound approximately 100m to the northwest (Burley 1996). If the panel is taken as an integrated unit, however, other of the motifs are difficult to reconcile, including two open bodied turtles, a lizard, a fish with arms, a human foot in outline, a pair of feet intaglio, two cupules and a depiction of an exotic female form. The only apparent superposition of images is the fish with arms overlying the possible snare.
Triangular Men on One Very Long Voyage 216 Figure 7. Houmale‘eia outlier images. Fig. 3 provides image locations.
Shane Egan and David V. Burley 217 The eastern panel occurs on the lowest exposed plane of beach rock (Figs 3 and 6). Panel imagery integrates ten dogs, three pairs of human feet in intaglio, a closed body turtle, possible clubs and three complete or partial anthropomorphs, with one of the latter possibly being a turtle/human transformation. The repetition and grouping of motifs again suggests the depiction of an event, though one hard to decipher. The clustering of dogs, clubs and a turtle could indicate the killing of a turtle with symbolic connotations in the human/turtle transformation image. Meaning, nevertheless, can be no more than speculation, and the panel could as easily constitute a series of unrelated depictions immersed in metaphorical symbolism.
The six outlier glyphs (Figs 3 and 7) include a human foot, isolated anthropomorphs to the south and west of the principal panels, and two anthropomorphs with associated
motifs to the east. One of the anthropomorphs is clearly a male with lengthened phallus. This image raises the question of gender identification based on the presence and characteristics of pecked dots between legs or at the body/leg intersection. In three cases the dot is cupule-size and potentially depicts a vulva, an interpretation sometimes offered in Hawaiian rock art studies (Lee and Stasack 1999). In most others, the dots could equally illustrate male or female genitalia.
THE CONTEXT FOR HOUMALE‘EIA (OR LACK THERE-OF) IN WEST
POLYNESIAN ROCK ARTFirst settled by Lapita “peoples” (after Kirch 1997), Tonga, with Samoa, Futuna and ‘Uvea, formed an ancestral Polynesian homeland within which a Polynesian cultural template developed over the subsequent millennium or longer. By c. AD 500–600, the template was transported eastward as the remainder of Polynesia was populated (Kirch 2000). Kirch and Green (2001) provide an extensive reconstruction of ancestral Polynesian culture employing historical linguistics, comparative ethnography and available archaeological data. A discussion of rock art in any form is significantly absent. A general paucity of recorded rock art sites in West Polynesia creates this situation. The few existing sites also cannot be positioned or characterised within an ancestral framework for petroglyph origins in East Polynesia (Wilson 1998: 181-82).
As part of the Bayard Dominick expedition to Tonga in 1920 and 1921, Thomas W. McKern spent seven months carrying out a survey and recording the myriad of archaeological site types found throughout the length of the archipelago. While understating the rarity of rock art, he (1929: 78-80) reports two petroglyphic types. Numerically the most dominant type is a diagonal and parallel line design carved on facing stones in several of the chiefly tombs at Lapaha, the 13th through 19th century capital of the chiefdom (Fig. 8). To our Triangular Men on One Very Long Voyage 218 Figure 8. Linear inscribed rock art on a facing stone at Langi Katoa, Lapaha, Tongatapu.
knowledge, this type of petroglyph is without comparative form in Oceanic rock art (Monnin and Sand 2004: 269-80, Wilson 2003), is exclusively related to tomb architecture and has an independent Tonga origin.
McKern’s other petroglyphic type occurs in a single site on the small island of Telekivava‘u and perhaps at another locale on Tonumea Island, both in the far south of the Ha‘apai Island Group. Although images were not illustrated, the Telekivava‘u site is specifically described as having figurative imagery, including a triangular body anthropomorph (McKern 1929: 78). A cadastral land surveyor, Larry Wordsworth, sketched the Telekivava‘u petroglyphs in 1957 (Wordsworth, per. comm. 2009). The site was situated on beach rock within the tidal zone and he was able to locate the images only after sunset employing a hurricane lamp for side-lighting. Later published by Palmer (1965), Wordsworth’s sketch depicts two anthropomorphic figures both identical to Shane Egan and David V. Burley 219 anthropomorphs from Houmale‘eia. (Fig. 9). An associated use of a cross with one of the Telekivava‘u anthropomorphs is also replicated at Houmale‘eia. The Houmale‘eia and Telekivava‘u artist(s) may be one and the same.
Numerous archaeological surveys throughout Tonga have been undertaken since McKern’s pioneering study, including systematic and intensive coverage of several of its widely dispersed islands (Burley 1998: 343-48); only one new rock art site has come to the fore. This occurred in 1991 when Burley (1994a) found an engraved foot on a facing stone in the royal tomb Mala‘e Lahi on ‘Uiha Island in Ha‘apai (Fig. 1). Also present at this tomb is a separate facing stone with the parallel line design reported by McKern. The foot was interpreted as an intentional marker symbolising the act of moemoe‘i, an ultimate sign of chiefly respect in traditional Tongan society through the touching of a high chief’s soles (Gifford 1929: 118). Foot petroglyphs are known across Oceania (Monnin and Sand 2004: 267), but they are especially abundant on the island of Hawai‘i (Lee and Stasack 1999: 186). They also are the third most abundant motif category at Houmale‘eia. The act of moemoe‘i may still be the intent for petroglyphic display of a foot at Mala‘e Lahi, but the foot’s origins can be re-evaluated. Rather than a direct execution, it seems possible if not probable that the petroglyph had been cut into a beach Figure 9. Telekivava‘u petroglyph panel sketched by Larry Wordsworth in
1957. Scale and north orientation are unrecorded. Reproduced with permission of Larry Wordsworth.
Triangular Men on One Very Long Voyage 220 rock surface that, later, was quarried for stone tomb construction. This type of quarry occurs in a number of locales in Ha‘apai, including one recently discovered to the front of the Houmale‘eia site.