CHAPTER 2 ENVIRONMENTAL AND ETHNOGCRAPIC BACKGROUND P. f/* Kirlh i~ ; & fMangareva is the collective name for Archaeological research in any Polynesian - AW a group of ten small "high" volcanic island benefits through close familiarity with the | islands encompassed within an encir- relevant natural history and ethnographic litera- cling barrier reef (23 07' S., 134"58' ture. MIany, aspects of the Mlangarevan env,iron- W), as well as the proper name of ment directlv affect the archaeological record the largest of these islands. Captain (such as dynamic sea levels and shoreline pro- James Wilson of the missionarv ship cesses), or aid in the interpretation of that record Dfuf encountering the islands on May 22, 1797 (such as the influence of soils or biotic resources (Hiroa 1953:47), named them after Admiral on settlement patterns). Archaeological interpre- Gambier.' In its proper historical usage, Gam- tation is likewise informed bv reference to the bier Islands applies strictlyT to this cluster of high ethnographic and ethnohistoric literature. In islands including Mangareva (Brigham 1900:96). Polhnesia, the cultures and societies documented Recently, however, the name Gambier Archi- at the time of European contact represented the pelago has been applied to a larger geographic "endpoints" of unbroken cultural sequences that entity including the nearby atoll of Temoe, as can be traced back to initial colonization. There- well as several atolls in the Actaeon Group fore, our use of ethnography in Polynesian ar- (Tenararo, Vahanga, Tenarunga, Matureivavao, chaeology falls wlithin what has been termed- and Mlarutea). It is in this broader geographic in North American usage-the "direct histori- sense that Iles Gamvier is used by the administra- cal approach" (Steward 1942; Strong 1953; tion of French Polynesia. Here we use the term Lightfoot 1995; (onte, in press), which is dis- Alangareva Islands to refer to the group of high tinct from the practice of "ethnographic anal- volcanic islands encompassed hx its barrier reef ogy". In this chapter, we summarize relevant as- and lagoon system, and Alangareta to refer to the pects of both NiIangarevan natural history and principal hig;h island, where the administrative ethnography as these relate to the archaeologi- center of Rikitea is situated. cal record. 17 ENVIRONMENTAL AND ETHNOGRAPHIC BACKGROUND NxvatuilA HISTORY o0' MANGARE\V\A breccias as xwell as many intrusive dikes (Brousse Beino- small and isolated, the Mlangareva Is- 1974; B3rousse and Guille 1974). The orientation lands have not been as thoroughly investigated of the numerous dikes exposed along the coasts by naturalists as other Eastern Polynesian archi- of Mangareva, Taravai, Agakauitai and Akamaru pelagoes. The Bishop Museum's 1934 suggests that a large central caldera wvas origi- Mlangarevran Expedition made important terres lnallv situated in the area now occupied by the trial biological collections, but these have been central lagoon, between these islands (Brousse only partly described or published (e.g., Kondo and Guile 1974, fig. 1). Our archaeological work 1962; Solem 1976; Zimmerman 1936). In the late has shown that many of these dikes were ex- 1 960s to earl, 1 970s, a multidisciplinary team of ploited as sources of raxv material for stone arti- researchers from the Service 7,lixte de Coentr6le facts, and flaked dikestone is abundant in some Biologique of France carried out a series of in- archaeological contexts as at the Nenega-ti and v!estigations for the Direction des CJentres Onemea sites (see C-hapter 6). As Brousse d'Expsrimentations N ucleaires, in conjunction (1974:178) points out, the chemical composition with the nuclear bomb tests at nearby Mururoa of the Mangarevan basalts is highly varied, with and Fano-ataufa atolls. The results of these in- at least three distinct groups ("tholeiites", "basalts and angaauf at()Ils Th reslts f tese n- lcalins" and "oea'ts)Welr(I96778 vestigations were published in two substantial oceanites"). Weisler (1996:76-78, fg6)collected and geochemicallyaaI-ed2 volumes in the series Cahiens du, Paci/fque bv the fi- an analyzed 26 Fondation Singer-Polignac; wxre have drawn rock samples from sevTen islands in Mlangareva, heavil from reportts in these volumes in prepar- as part of his program of tracing prehistoric ing this summaryT of Mangarevan natural histor. exchange between the MNangareva and Pitcairn islands. The variability in Mangarevan rock GE-OI ACY} ,4AN!) GXE( )AI( )RPHOL (- GY geochemistry- has olbvious implications for ar- As is typical of true "oceanic" islands situ- chaeolo gical efforts to characterize and source ated on the Pacific Plate, Mlangareva has a "hot artifacts made of volcanic rocks. spot" origin, on the same volcanic alignment with The high islands and lagoon are protected Pitcairn Island to the southeast (Munschy et al. on the xvest, north, and east by the extensive 1998). Radiometric (1K-Ar) dating of volcanic barrier reef syTstem; however, the reef is sub- rocks from the Mlangareva Islands have yielded merged to the south, permitting storm swells to ages of 4.77-5.98 Mla (Bellon 1974), and 5.66- enter the lagoon from that direction. Fringing 6.26 Mla (Guillou et al. 1994). With this age of reefs are found along tlhe coasts of the high is- rou hlv six million years, Nangareva has ml- lands, although the southern coastlines tend to grated in a northwesterly direction from its be more exposed and in places are marked by original position on the Pitcairn hot spot, in formidable sea cliffs. This is the case with the process undergoing both subsidence and Agakauitai, for example, which has a fringing reef extensive subaerial erosion. The highly dissected on the northwest, but exposed cliffs on the south islands encompassed within the barrier reef and and east. Likewise, Makaroa and Kamaka islands lagoon are thus what remains of a once much have fringing reefs and small sand beaches only more extensive hicyh island (Fig. 2.1). Table 2.1 on their more cted, shores. lists these ten volcanic islands in order of size, large island of Nangareva, being the most pro- giving their respectivTe areas in square kilometers, tected by the barrier reefs, has extensive fringing and their maximum heights. reefs and sand flats surrounding it. lDeep, pro- The igneous rocks making up the MSangareva tected bays are found along the northwestern islands are varied in ty pe and geochemical com- coasts of MIangareva and Tarav ai islands. These position, and include blockl lavas and pyTroclastic variable coastal conditions profoundlzy influence 18 ARCHAEOLOGICAL INVESTIGATIONS IN THE MANGAREVA ISLANDs, FRENCH POLYNESIA . . . . . .. . . . . . . . . . . . . . ... . .. . . . . .. . . . . . . . . . . . . . .. . . . .. . . . . .. . . . .. . . . .. .. . .. . . . . . . . .. ...... .. .......... .......... . .. ....... . ... ....... 7-7 7 w. . ..... sw ? ..... ...... 7: ......... . .. ..... ......... ....... ... ....... . ..... ...... ............. X , .. . ............ ..... ...... ..... ...... .. ............... .. .. ...... .. ............. ....... ...... .......... .... .. ... ............. ..... ............. .............. .................. .. . . .. .. . . . . 7 : 7 ............. . ........... ......... ..... .................. . ...... .... .. ..... .. . . .. .. . . .. .. . . . . . . . . .. . . . . . . . . . . . . . -X :-X ............. .. .... . .......... ........................ ................ ...... ............ :.: . - - : :. :-: s., a.:?:: ? ? :.;.:.:.:.:....?..?:..?.;...?.I..?. _ .? ...- ...... .......... .. . . ... .... . . . . . . .. . . . . .. . . .. ..... ..... . ...... . . ....... `:-X . ............ .. . .. .. ... ..... .. . . . ... ....... .... .. . . . ......... . r: :: ....... . .. .......... .. .. ................... ... ... ............ ...... .. ..... ............... .. -X .................. MOTU TEAVAONE m;rm 'MOTU TEPAPURF:?' ............ .. I. I ... :. " '. . - !.::...:. - .:::? ...... . ........ .............. .... .... ... . ..... ............. rw:- m ma::: :: ....... .......... ... . ...... :: ?;:::: ?.': - 7 ..... .... . ............. ........ . ..... ........ ... ... v - - .... ....... X . . ...... a . :7.::: .... .... ......... - :7- - ,-: X. X. .. ............. .. .. ........... . ............................. .................. . ....... .... ...... ..... . 77 _X, .... .. ........... X ...... ... ........ .. .... .. ... .... ... . . ......... ,MOTU PUAU m U "iV.", Fr: .......... . ............... ......... ............ .... w?x :; -XX-:v-X-::: :i .... ......... .... ..... ..... . .......... .............. ........ ..... . ............... .............................. ........... . .. .. ................. .......... .... ........... ... ... ..... . .. . ............... . ........... .......... .......... ....... . : ... ..... ... ... . . . . .. . . ..... . . . . . .. . . .. .. . . ....................... ............ .......... ............ ............................ ............................ .... ......................... ............. . .................. .... :. ? ?. .. . ..... ..... ........ ---- -- ............. . ..... ......... ........ 7::: ............ ...... X-X ?X : .......... ............. .... .... ... ..... ... .......... ... ................. ............................ ........................... ................... ............. ............. V. -X-X-:- X.: . ...... .............. ..... . ..... . .......... .......... ... .. .... ....... ..... ................. .. ..... ........... .................. .. . ......... ..... ..... . .. ................ . ............. .. ................ . ........... ............ ....................... ... ........................ ............ ........... . . . .. . . . . . . ... ................ ..... ... ................ .. . ..... ..... .. . .. .............. .. .. ...... ....... ......... . .... ... ....... .............. v X . .................. . . ...... .......... ...... ... ... . ........ ............. ..................... . ... . ......... ... . ..... ......... .. ........... ......... ........... .............. ........................... ....... ......... ............................. .............................. ... .. .......... ... ... .. ........ . .. ....... ........... . . MOTU TENOKO ............ -7 7 X .......... .... .... .... ... .. .. -:7 Akaputu .......... ........... ........... . .......... .... ....... . . ..... .... .... .... ........... .......... A p e a k a v a i, .. ........ d. .. ...... ................ .... .......... . . ........ ...... ........ .. ........ ...................... ?. . - .1-1- 1........... .:.:: ma ................. . ......... Gatavake,,, M ANG AREVA Q? ....... .......... ........... v ma amwoo::: ??: A tia oa Rikitea .... ....... .. ..... ... ........... 'AUKENA ............. Auorotini (441 ryi) ........... A titu iti ........ .......... ....... .... W, 7: ............. CTA RA VA I ............ ............ . msr . . ......... X'X ........... ............... mx.. a MOTU TE AGAKAUITAI . .. ......... . ...... M EIKIRO .......... . ..... ... . Tokuruo Bank ..... ... .... 'AKAMARU ........ .... a . .... ....... ............. ...... .......... .............. .......... ..... . ......... ...... .......... ............... ............. .. ... ............. .... ........ w. KOUAKU .......... ............ .. .... . ...... ............... ......... . ...... ....... ...... . . . ................. . . .................... .. .... ... .. .............. ............... ... .. MOTU TEIKU ........... ............. .. . ................. . . ........ .. . ........... ... .... ...... OM ................. . ..... ... .. . .. ..... -,MAKAROA ........ ..... .. ........... . ...... .... X!; ...................... . .. . . . . . . . . . . . . . . .... ..... ....... .......... . .......... . .. . .... ....... ............. ............. ....... .............. X X. ... ...... .. ..... ............ a : :, - :::::::-:::-:? . ......... . .................. MANUI ....... .... .... .............. ........... ............ .......... .......... . .... ... .... ............ .............. ...... ... .. . .... ..... . ...... Xo "' X . ......... .......... ..................... KAMAKA .............. Ma n areva lsia n ds X X.: X .... ........ . .... ...... ............ ............ 0-1 5m depth TN r: a . . . . . . . . . . . . . . . a :X.: . . . . . . . . . . . . . . . . . . . . . .... . . . . . . . . . . . . . . . . . . . . . . . . . . . ... M N -A X F . X a 15-60m depth . . . . . . . . . . a a . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 a . . . . . . . . . . . . . . . . . . . . . . . . . . . .... . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . ocean . . ... . . . . . . . . . . . . . . . . . . . . . . . . . . . .... . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . N. .... .. . . . . u 1 2 3 4 5 km . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . . . . . . . . . . . . . . . . . . . . . . . 7-- x W.: 4.. .. . . . . . ... . . . . . . . . . . . . . . . a . a a X.: . . . . . . . . . . . FIGURE 2.1 Map of the Mangareva Islands showing the principal high islands and motu of the barrier reef, and submarine topography of the lagoon. (Map after Atlas de /a Polyn6sie Fronqaise, ORSTOM, 1993, plate 16). 19 ENVIRONMENTAL AND ETHNOGRAPHIC BACKGROUND TABLE 2.1 Geographic characteristics of the volcanic islands of Mangareva.* Island Area Highest Comments (kM2) point (m) Mangareva 14 441 Largest island and administrative center (Rikitea Village); several large valleys and deep, protected bays. Taravai 5.3 250 One village, nearly abandoned today. Three large bays on the west coast; large valley on the east coast. Akamaru 2 246 Large coastal plain on the north; one large valley on the west (Tokani). Aukena 1.5 198 Two peaks with narrow isthmus. Narrow coastal plains. Agakauitai 0.7 139 Two small valleys on the west side; cliff bound on the south, east, and north. Kamaka 0.5 166 Privately owned by the Reasin family; beach ridge on the north, rockshelters. Makaroa 0.2 136 Small valley and beach ridge on the north; otherwise cliff bound. Mekiro 0.075 58 Small, cliff bound islet. Manui 0.070 54 Small, cliff bound islet. Makapu 0.065 |-50 Small, cliff bound islet. *Areas and elevations after Brousse (1974). at Gatavake on M\angareva, or Aganui on As reported in Chapter 3, we also noted such Taraval); others have small drainage channels that stone structures in the intertidal zone at Atituiti- flow after heavy rains; small springs or seeps Raro on Mlangareva Island, and along the coasts emanate at the base of other valleys. Cliffs oc- of Aukena and Akamaru islands. Other obser- cur commonly, with shallow rockshelters formed vations, such as the frequent presence of wvave- where strata of softer breccia have been eroded cut banks and active erosion of beach ridges, also more deeply than the intervening dense lavas. reinforce our view that the NMangareva Islands An important issue for archaeology is the are actively undergoing a phase of relative trans- matter of relative sea-levels and shoreline dvnam- gression of sea level, with erosion of archaeo- ics. Brousse et al. discuss the geomorphological logical sites and deposits. This is a topic that de- evTidence for subsidence over the longer term of serves particular geoarchaeological investigation. geological history: ". .. nous pensons que la sub- sidence a et le phenomene qui a provToque la Cl,Am disparition de la plus grande partie de l'edifice At 23" S. under the Tropic of Capricorn, the vTolcanique dont nous n'apercevTons aujourd'huique MIangareva Islands have a somewhat cooler cli- le sommet en grande partie demantele"' (1974:90). mate than the Society, or Mtarquesas archipela- They do not comment on xvhether there is ac- goes (C h&vre 1974:144). The trade winds blow ti've subsidence at the present time. However, predominantly, from the east. The annual aver- Orliac (2002, 2003) reports a number of archaeo- age temperature is about 24( C., with the period logical features, such as stone walls and platforms from about May to October being somewhat along with cultural deposits, which are presently cooler. Average annual precipitation ranges be- in the subtidal zone at Gatavake and other bays. tween about 1,400)-1,90)(: mm, with the heaviest He contends that these are evidence of as much rainfall concentrated in DGecember-January, al- as 50 cm of subsidence over the past 800 years. though the differences between the rainiest 20 ARCHAEOLOGICAL INVESTIGATIONS IN THE MANGAREVA ISLANDS, FRENCH POLYNESIA .~~~~~~~~~~~~~~~~~~~~~~~.~~~~~~~~~w .............~~~~~~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~~~~~~~~~~ -:-m-;- 'wf .*.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. ~ ~ ~ ~ ~ ~ ~~ ~ ~ ~~~~~~~~~~~~~~~~~ . . ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ..... FIGURE 2.2 The Mangareva Islands are characterized by small high islands clustered within a large lagoon and barrier reef system. In the foreground is Aukena Island, with Akamaru, Kamaka, and Makaroa islands in the distance. Photo by P.V. Kirch fl | , . .-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. ........ .. .. FIGURE 2.3 The topogareaph oflnd the Manarsevandbysml high islands istpialyseeswtherassihi covrgedsoe bkroke bysranks of cle ifstne Photo by P.V. Kirch . ........, . ....*D ,;. %tY0 FIGURE~~~~~~~~~~~~X 2. h oorpyo h agrea ihilnsi yial sep ihg1scvrdsoe broken by rsnks ~of lfs.Poo yPV.Krh 21 ENVIRONMENTAL AND ETHNOGRAPHIC BACKGROUND month (December, 4206.2 mm) and the driest grasslands. Finally, the hydromorphic soils- month (August, 131.6 mm) are not huge. With while restricted in geographic distribution and this climatic regime, Mangareva is well suited to extent have considerable economic signifi- the cultivation of the suite of tropical cultigens cance, for these were the main soils used for irri- imported by, the Polynesians throughout the cen- gated taro cultivation. Theyr have abundant or- tral Pacific region, including taro, breadfruit, co- ganic matter, a pH of around 6, and high avail- conut, bananas, and other crops. able nutrients, including phosphorus. On S()ILS IMangareva, the largest area of hydromorphic soils is found at Rikitea and along the coastal MIore limiting for the development of tradi- plain from Atituiti-Raro to Ganoha (very small tional horticultural syrstems are the Mangarevan areas of hydromorphic soils in other valleys do soils. Tercinier (1974) classified, inventoried, and not appear to be mapped b Tercinier [1974: mapped the soils of Mangareva Island, which carte]). As we discuss further in Chapter 3, the presumablyT encapsulates the range of edaphic large area of hydromorphic soil at Rikitea is variability in the islands. He defines seven major doubtless a major reason for the early settlement groups: (1) "sols mineraux bruts d'erosion" and continued sociopolitical dominance of this (lithosols); (2) "sols peu evolues", including col- localit}; as it supported the largest expanse of luvial and alluvial soils; (3) "sols vertiques irrigated taro land in the archipelago. (vertisols); (4) "sols calcomagnesimorphes"; (5) "sols brunifies"; (6) "sols ferrallitiques"; and (7) F1i )ORA ANXi) V/IGET4TFIOXN Al TIJTRXS "sols hydromorphes". The lithosols are found Huguenin (1974) provides a checklist of mainly on the flanks of Mlt. Duff and Mlt. about 200 species of ferns and higher plants in MIokoto and have no agricultural significance. the Mlangareva Islands, but there appears to be The alluvial soils of category (2) are among the no definitive study of vegetation patterns. From most important for the traditional Mangarevan our own observations, the most striking aspects economi, as these comprise most of the lower of the vegetation are: (1) the absence of native elevation slopes of the principal valleys, as well forests on the steeper slopes and ridges domi- as alluvial in-filling of the valley floors. These nated by degraded fernlands and canelands; and, soils tend to have high organic matter, a slightly (2) the strongly anthropogenic character of the acid pH, and are relatively high in available nu- vegetation in the valley bottoms and coastal trients, such as phosphorus (Tercinier 1974:368). plains, dominated by economically useful plants, The steeper windward slopes (above the collu- many of them PolNnesian introductions. We have vial fans) are dominated by vertisols and are alreadv referred to the strong associations be- stronglyv associated with extensive stands of tween fernlands of Dicranoptenis (Gleichenia) /ineanis Aliscanthus cane. The "calcomagnesimorphes" and the ferralitic soils, and betwveen canelands soils are a category of carbonate soils formed of AIiscanths/floidu/lus and the vertisols (Fig. 2.4). on coral sand sediments around the littoral These vegetation associations are maintained by fringes of the island, especiall at the mouths of fire, and the low nutrient status of the soils in- the major embayments (such as Rikitea, hibits secondary reg-rowth. That the degraded (;atavake, or Atiaoa). The ferralitic soils occur fernland and canelandi vegretation dominating the extensively on the mountain crests and the lee- TMangarevan hi,h islands was not an artifact of wvard slo)pes (where they replace the vertisols of po)st-European contact changes in land use is the windward slopes); exchangeable nutrients are made clear by this observation of C aptain Wil- extremelyT limited, and these soils are essentially son of the Du!/' in 1797: worthless for cultivTation. The ferralitic soils tend The tops of thle hills, to about half waya down, are to) be associated wvith Dicranopteris (G/eichve;iia) chiefly covTered with sun-burnt grass; and in some 22 ARCHAEOLOGICAL INVESTIGATIONS IN THE MANGAREVA ISLANDS, FRENCH POLYNESIA FIGURE 2.4 The vegetation of the high islands is highly anthropogenic. Coastal areas are dominated by Pandanus odoratissimus and Thespesia populnea trees, with Hibiscus tiliaceus on the narrow coastal plains and in the shallow valleys. The higher slopes and ridges are dominated by grasslands of Miscanthus floridulus. Photo by P.V. Kirch. places there- are 'spots ot reddislh s(il, as~ on the- (Gcoro nuczfe;a), breadfruit (Artoca;pus a/ti/is), Ta- midldle gro undls ot ( tahcitc (1 799) 118). hitian chestnut (Inocaipus Ata>ifrrus), candlenut In our opinio)n the expanses of Dlicianopteris (A/euriter mo/uccana), and iPi apple (Spondias du/cis). and Al iscau?thus which have covered the higher The narrowv upper valley watercourses tend to elevTations of the MIan,rarevan high islands in his- be chokied with dense stands o,f Hibi'scus tili'aceusr, toric times were not a natural successional state, while alono the coastal strand one commo)nly but rather an artifact of human land use prac- encounters P'auda;rnr tectorius, Hernanzdia pe/tata, tices during prehistor!-. We wxould hr pothesize c a/opfr'//urns znop/)'y//u;/, BSa1r;7ugtonza asiatica, C'ordia that these fern- and canelands developed in re- IWbcoi data, Terminalia cazt9ppa, and T'hespesiapopu/nea. sponse to burning and fotest cleaiance on slopes Plants fo)und in the understory of coastal and xvith ovld, nutrient-poor soils, much as xxvas the vTalleyT second groxvth, and which w^ere introduced case in MVangaia (Kirch 1996, 1 997b)) Howvex ci to the islands bxT Polynesians, include bananas this hypothesis needs to be tested through pa- (Alu.sa fe/i [Alu t;a/zrnusaj and Eumnusa hybrids), leobotanical investigations, such as anaix sis o)f Pol nesian arrowroot (i2icca /eoutopeta/oidess), ti microfloral remains (pollen, o)pal phv7toliths) from (C.oidy/me Ii utzcosa), kape QA/ocasia rnzacro0rrhiga), and sedimentaryr contexts. non?o (MVorinda cil7O/ia). The valley bottoms and coastal phains are It is difficult to overly stress the huge impact dominated by a range of economicaibr impor- that human activities have had in shaping the tant plants. Principal tree crops include coconut historically known flora and vegetation patterns 23 ENVIRONMENTAL AND ETHNOGRAPHIC BACKGROUND of Niangareva. Not a single endemic species of cers and carapace fragments that we have tenta- higher plant is recorded in Huguenin's checklist tivelvt identified as being of this taxon (see Chap- (1974). Harold St. john, botanist of the 1934 ter 5). Aside from lbirds, the only v,ertebrates listed Bishop Mluseum MIangarevan Expedition, re- by C(ochereau (1 974:516-17) are three species of corded his utter disappointment with the botani- lizard, including the PolyNnesian-dispersed gecko cal possibilities of Mlangareva in a single line of (Gehira oceanica), a freshwater eel found in taro his report: "Mangareva Islands are desolated; pondfields (Anig/amillegastoz a,eandtheEuropean their natural flora is more completelyT extermi- introduced Rattus rattus (which has evidently nated than that of any- other part of the world eliminated the Polynesian-introduced Rattus that 1 have seen" (1935:57). The same view is exulans). presented bv the 1934 Expedition's leader and In view of this extremely impoverished ter- malacologist C. MIonta?gue Cooke, Jr., who wrote restrial fauna, birds provided the only significant that "all the endemic forests have disappeared... terrestrial resource from the viewpoint of sub- except on the precipitous southern slope of sistence economy. Lacan and Mlougin (1974) re- MIount MIokoto, xTwhere some of our party found view the extant avifauna, listing 23 species in total a small remnant of native forest near the base (including the domestic chicken, Gallus domesticus), of the cliff. A few scattered native shrubs and the list being heavily dominated by sea birds. small trees were growing on the ledges above" There is a native kingfisher (Halcyongambieni) of (1935:41).3 a species found also in the Tuamotus; the only other land birds are a reed-warbler (listed as Ti-,RRI-,S`FRL4L FALTI 1-A AND[ RPI SOL.TRCES Coapoea .q I T-RRESTRiALFAu AN) RE RCE CO;opoderas caffrUa, although this species is sup- Cochereau (1974) inventories the terrestrial posed to be endemic to Tahiti and Mlo'orea) and fauna of Mangareva, which is dominated by in- the common rock dove (Columiba lit'ia). Lacan and vertebrates, particularly insects, among which one Mlougin (1974:537) stress the uneven geographic does find a number of endemic species (e.g., distribution of seabirds among the higrh islands Zimmerman 1936). Terrestrial molluscs are to- and coral islets (motu), especiallv the nesting and day represented by only six taxa, three of xvhich reproducing populations, which are heavily con- are widely dispersed pulmonates thought to have centrated on three small, high islands in the been transported inadvertently by the Polynesians southern part of the lagoon (Nlakaroar lManui, (Tornatel/inops mariabilis, Elasmias apertum, and and MIotu Teiku). They regard this distribution Lamellidea obloqoa; see Kirch [19984:137j). Subfossil as directly related to the relative lack of human pres- deposits (including our owrn excavatio)ns), ho)w- ence on these small and difficult-to-access islets. ever, have yielded other taxa such as several en- Based on zooarchaeological and paleonto- demic genera and species of endodontids (Solem logical studies on other Polynesian islands 1976) wvhich are evidently now extinct. We dis- (Steadman 1989, 1995, 1 997a), one mnay predict cuss these further in Chapter 5. The only terres- that the Mlangareva Islands originally had a more trial crustacean listed byr Cochereau (1974:489) diverse avTifauna, includin T RCIxIzAR!ox Mangarevan material culture, based largrely on AND L D NE) {IE 19th century accounts and on examination of M'tangarevan social organization was com- such specimens as survive in museum collections. plex, involving cross-cutting categories based on He classified the stone adzes available to him for male primogeniture, on affiliation through de- study into three main ty pes: Type 1, quadrangu- scent or adoption with an eponyTmous land-hold- lar section; Type 2, quadrangular wvith rounded ing group, on success or failure in wars over land, edges and reduced polls; and Type 3, thick, long and on acquired statuses (such as warriors, ex- 27 ENVIRONMENTAL AND ETHNOGRAPHIC BACKGROUND perts, or priests). People were categorized as ei- (Hiroa 1938a:151), the child was taken to a ther elites or "nobles" (tq,o7tWi), or as commoners "house of seclusion" situated on a high flanking ('urumnaniu) based on birth, the latter performing ridge of Auorotini, wrhere he wvould be cared for the bulk of subsistence labor. Hiroa writes that until he xvas about twelve or fourteen years of age. "the nobles had power (ao), land (kaigda), a supe- Emorn (1939:22-23) refers to these houses as the rior type of house ('are), a tribe or people (u), "roval nurseries" and describes two stone pave- and a freshwater spring (ui)" (1 938a: 144). WUhie ments that mark these sites (see Chapter 3). Fol- commoners could not become togo'iti, they could lowing his descent from the mountain nursery, through acquisition of particular skills become at puberty, the voung chief would be installed in specialists, such as warriors ('aretoaq), master crafts- "a royal residence at Mlarau-tagaroa" (Hiroa men (tu'ziga), or priests (taura). Hiroa writes also 1938a:152). Hiroa notes that the house of the of a "middle class" (pakaora), which included 'aka;ikiwas larger and better constructed than those junior lines of noble families, and of common- of other chiefs and that it contained a stone bench ers xwho had been elevated to that status throu,h upon which the chief seated himself. "grants of land for services in xwar" (1938a:146). Beechey (1831:193) offers a description of Those who held large blocks of land were re- Mlangarevan houses prior to the major changes ferred to as riagatira. in house type initiated by Laval, and draws a clear The Mlangareva Islands seem not to have distinction between the small houses of the com- been politically integrated under a single ruler mon people ("in length from eight or ten feet to but rather were divided into several independent fifteen") and "the larger houses of the areghe and frequently warring polities, each encompass- lchiefs]." This latter dwelling xvas described in ing a principal island or in the case of NMangareva (greater detail: Island, one of the two districts into which this The large house, or that of the ar-eghe, was about larger island was subdivided (Taku and Rikitea). thirty-nine feet in length by eighteen or twenty in The high chief who headed up such a polity was width; the pitch of the roof was about twenty-five called the 'akariki, an interesting variant of the feet in height, and that of the perpendicular sides ca'led the .akar, an interesting vanantofthe of the house about ten feet; but these dimensions Proto-Polynesian term for chief (PPN *qaiki), w o ; ' w~~~~~~~~~~~~xere obtained bv estimation o)nlv, the nativTes formed by, combining the causative prefix 'aka appearing to have an objection to our pacing the (PPN haka) with the term for chief (a;iki). Hiroa ground for the purpose of measurement. The so Mth reads considerable historical significance into this side of the house was left open . . . On that part of linguistic construction, suggesting that it "implies the house where the side was deficient, there was a that the ariki position had to be created in foundation for the wall about three feet in height Mtangareva at some period when ahereditarv ariki thrown up, composed of large blocks of coral, shaped in a very workmanlike styTle, similar to those did not exist" (1 938a:1 51). However, the term is mentioned by Cook at the Friendly Islands, and well cognate with MIarquesan haka'iki (a parallel con- put together: it stood about three feet within the struction), and following Fischer (2001) it mayT outer part of the roofing, and served as a seat for simply be a loan word into Mangarevan from the chiefs as well as for manv others (Beechey Mtarquesan. 1831193-94). The 'akariki of Rikitea, who at times also Land tenure was a complex matter in held sway over the whole of Mlangareva Island, Mlangareva, involving both hereditary rights of was supposed to be born on the marae of Te descent groups ("tribes") to particular ancestral K