Seikel 2011 Mortuary Contexts and Social Structure at Nan Madol

Mortuary Contexts and Social Structure at Nan Madol, Pohnpei

Katherine Seikel Department of Archaeology and Natural History, School of Culture, History & Language, College of Asia-Paciﬁc, Australian National University, Acton, Australia Capital Territory, Australia

ABSTRACT

Archaeological studies of mortuary practices contribute to discussions of prehistoric social structure. Relatively little work has focused on mortuary practices in Micronesia, particularly those associated with societies that built the monumental earthworks of Palau, the Latte sets in the Mariana Islands, and the monumental centers of Leluh on Kos- rae and Nan Madol on Pohnpei. This study utilizes multiple lines of evidence to consider mortuary contexts and their relationship to social structure at the site of Nan Madol. The mortuary context is interpreted as reflecting ascribed status in a stratified hierarchical system consistent with traditional histories. The study provides a base line for future studies of Pohnpeian mortuary practices and adds to archaeological knowledge of complex societies in Micronesia and the Pacific.

Keywords Pohnpei, Nan Madol, monumental architecture, mortuary contexts

Data from burial practices and be- ology (e.g., Binford 1971; Huntington and liefs about death were of central im- Metcalf 1979; Rakita et al. 2005; Saxe 1970; portance to the early development Silverman and Small 2002; Tainter 1978) be- of the study of human evolution, of causeburialcontextsembodymultiplesocial the rise of ancient civilization, and processes (Ucko 1969:276). In this respect of cultural and social institutions. mortuary studies are important for under- (Huntington and Metcalf 1979:5) standing social structure in both the present and the distant past. Detailed archaeological The study of mortuary practices has long discussions of mortuary practices in the Pa- been a focus in anthropology and archae- ciﬁc have been largely limited to the last 10

Received 6 May 2010; accepted 8 December 2010. Address correspondence to Katherine Seikel, Department of Archaeology and Natural History, School of Culture, History & Language, College of Asia-Paciﬁc, Australian National University, Coombs Building 9, Fellows Road, Acton ACT 0200, Australia. E-mail: [email protected]

442 Mortuary Contexts and Social Structure

years (Anton and Steadman 2003; Fitzpatrick shifts in power. Oral traditions divide Pohn- and Nelson 2008; Parke 1998; Pietrusewsky peian history into four periods: the Period et al. 2007; Scott et al. 2010; Valentin 2003; of Building (c. AD 1–1000), the Saudeleur Valentin et al. 2010). These studies enhance Period (c. AD 1000–1500), the Nahnmwarki our understanding of the development of so- Period (c. AD 1500–1820), and Contact cieties and social structure across the Pacific. Period (c. AD 1820 onward) (Hanlon 1988). Monumental architecture and large- The first period is characterized by the scale constructions are at the center of stud- creation of the island and the establishment ies dealing with settlement patterns and soci- of matrilineages, which form the basis of etal development in the Pacific (Ayres n.d.b, the Pohnpeian clan system (Bernart 1977; 2002; Clark et al. 2008; Graves 1986; Kirch Fischer et al. 1977; Hanlon 1988). The 1980, 1990; Liston 2009). Although Microne- Saudeleur Period began with the arrival of sia contains a variety of major constructions two foreign brothers who are attributed (i.e., Palauan earthworks, Leluh on Kosrae, with starting monumental construction at stone money on Yap, Latte foundations on Nan Madol and the founding of the Saudeleur the Marianas, and Nan Madol on Pohnpei) dynasty (Bernart 1977; Hambruch 1936; they are generally under-represented in re- Hanlon 1988). During this period established cent monographs on the development of local clans were incorporated under the complex societies in the Pacific (e.g., Kirch Saudeleur as fiefs, extending Saudeleur rule and Rallu 2008; Lilley 2006). This under- over Pohnpei (Bernart 1977; Hanlon 1988). representation can be attributed to a lack of Although the Saudeleurs were initially familiarity with research conducted in Mi- welcomed by the people of Pohnpei, by the cronesia during the past few decades (e.g., time of the last Saudeleur they were discon- Rainbird 2004). Further research building tent with the increasing amount of tribute upon past work will contribute to our under- they were required to provide (Bernart standing of complex societies in Micronesia 1977). The arrival of Isohkelekel and the and other parts of the Pacific. overthrow of the Saudeleur dynasty marked One of the most striking Oceanic sites the beginning of the Nahnmwarki Period containing monumental architecture is the (Hanlon 1988). Isohkelekel established the megalithic complex of Nan Madol on Pohn- Nahnmwarki system, which created a new pei. At Nan Madol, columnar basalt was used hierarchy of chiefly titles and returned socio- to build islets, seawalls, large burial com- political authority to the clans. Isohkelekel plexes (lolong), and foundations for residen- divided Pohnpei into three independent tial and ritual constructions (Figure 1). Nan polities (wehi), each of which was governed Madol’s size and complexity provide chal- through a dual title system (Hanlon 1988). lenges to mapping and interpreting the site Within the title system Nahnmwarki was as a whole. Previous studies (Athens 1980, the highest rank in the primary line of 1984, 2007; Ayres n.d.b, n.d.c; Bath and titles. The Nahnmwarki Period marked the Athens 1990) focus on specific areas of the decentralization of socio-political power, site, architectural types, or artifact distribu- though Nan Madol continued to be occupied tions. This study contributes to previous in- (Bernart 1977; Hambruch 1936). Nan Madol vestigationsthroughapreliminaryanalysisof was abandoned by the early Contact Period mortuary contexts at Nan Madol. It provides but continued to be used as a ritual site into a framework for discussions of social hier- the nineteenth century (Gulick 1859). archy in relation mortuary contexts utilizing Nan Madol is located on a reef flat on multiple archaeological data-sets. southeast Pohnpei, a volcanic high island in the Federated States of Micronesia (Figure 2). NAN MADOL: A POHNPEIAN Itiscomprisedofapproximately100artificial SOCIOPOLITICAL CENTER islets spanning approximately 80 hectares (Bath and Athens 1990). Islets were built The traditional history of Pohnpei is marked on a foundation of coral and basalt boul- by multiple in-migrations and socio-political ders with coral rubble islet fill; columnar

JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 443 Katherine Seikel

Figure 1. Architectural photographs of Nahndauwas: A. View of west facing wall; B. Main entrance to Nahndauwas. Note photographs correspond to plates 40 and 41 from Saxe et al. 1980.

basalt was used in the construction of archi- religious center into the 1800s (Gulick 1859; tectural features (header-stretcher construc- Hanlon 1988). tion) and in some cases was incorporated The wide range of identified burial in islet walls. Radiocarbon dates and eth- types on Pohnpei indicates a variety of mor- nohistoric accounts suggest that the build- tuary practices. These include tombs (lo- ing and occupation of these islets occurred long), burial platforms, burial mounds, bone from approximately AD 1000 into the mid-to- caches, and sub-surface interments (Ayres late 1700s (Athens 1990, 2007; Ayres 2002; and Mauricio 1997:9–13). Lolong are the Bath and Athens 1990), though there is ev- most visible mortuary features on the is- idence of occupation on the reef flats as land with walls up to five meters high, early as 2,000 years ago (Athens 1990; Ayres but available radiocarbon dates (Ayres et al. n.d.c). By the AD 1700s the Nahnmwarki of 1981) indicate that they do not appear in Madolenihmw moved his primary residence the sequence until after AD 1200. This coin- to another part of the district, possibly as cides temporally with the major construc- the result of disease or a typhoon (Davidson tion phase at Nan Madol (Athens 2007; 1967:90). After the Nahnmwarki relocated, Ayres and Scheller 2002). Regardless of its the site ceased to function as a political cen- relatively late date of construction, Nan ter; however ethno-historical accounts sug- Madol contains all burial types except burial gest that Nan Madol continued to be used as a mounds. As Bath and Athens (1990:279)

444 VOLUME 6 • ISSUE 3 • 2011 Mortuary Contexts and Social Structure

Figure 2. Map of Micronesia showing the location of Pohnpei and inset of Pohnpei showing the location of Nan Madol. point out, “the islets comprising the outer INVESTIGATING SOCIAL STRUCTURE wall of Nan Madol contain the majority of tombs and/or mortuary features, and the Given the size and complexity of Nan Madol, fewest platforms/foundation features.” Even this study utilizes relatively small data-sets thoughtheseawallofthesitemaycontainthe to test a range of hypotheses about so- majority of burials in architectural features, cial structure at the site. Understanding the identiﬁcation of the seawall as the pri- mortuary contexts is best approached us- mary burial location is premature as it over- ing multiple data-sets (i.e., architecture, ar- looks the potential presence of less visible tifacts, paleodemographics) because these forms of burial at Nan Madol. contexts are the culmination of many social

JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 445 Katherine Seikel

processes (Ucko 1969:276). Discussions of Visibility is important for studies of hi- multiple variables from archaeological data- erarchy because the concept of vertical sets will allow for greater understanding of hierarchy suggests greater status is associ- social structure in relation to mortuary prac- ated with large or monumental structures tices at Nan Madol. Before addressing how over smaller structures or features (Christian the individual variables will be approached n.d.:30). Ashmore and Geller (2005:92) it is important to briefly outline a few ar- note that high-status burials tend to be eas that impact the overall approach to the more formalized, which suggests that bone data. caches are more likely to be associated with Social structure is generally categorized high-status burial practices since they con- as egalitarian, ranked, or stratified. Earle stitute another stage in communal mortuary (1987) notes that there are many common- practices. alities in the social structure of ranked and A final issue to consider is whether social stratified societies, an observation which has structure may be obscured in mortuary as- sparked considerable debate in archaeology sociations due to temporal (i.e., elaboration on how each category should be defined of burial architecture) or taphonomic (i.e., (Cordy 1985; Earle 1987; Paynter 1989; Pee- secondary burial outside of tombs) factors bles and Kus 1977). Rather than entering into (Tainter 1978:121). It is necessary to ac- this or similar debates, which have produced knowledge that identified burial types at Nan no clear archaeological distinction between Madol may result in part from chronological ranked and stratified societies, this study will change in burial practices, rather than distinguish between egalitarian and hierar- resulting from social structure alone. Even chical social structure. Social hierarchy will though this is likely, chronological data on be discussed in relation to achieved versus mortuary contexts at Nan Madol is currently ascribed sociopolitical standing. Egalitarian unavailable. Irrespective of this, mortuary societies tend to have minimal social distinc- patterning in the archaeological record can tions, which, if they do exist, are achieved still give us insight into the prehistoric social on merit. In societies with hierarchical sys- structure. tems social status is typically achieved on The questions listed below frame the ap- merit or ascribed at birth. For example, proach used with each data-set discussed in Hughes (1982:6) suggested that high-status this paper. titles were ascribed whereas lower status titles may have been achieved during the Nah- nmwarki Period on Pohnpei. Question 1: Are Differences in Mortuary As Tainter (1978:117) explains, the pri- Architecture Related to Social Hierarchy? mary purpose of “classifying mortuary data is to isolate clusters of burials which then can Graves (1986:149) suggested that stone be interpreted as socially distinctive.” Classi- constructions, being relatively permanent, fication is also important because burial rites reflect the social position of community vary in relation to social status (Huntington members associated with them; therefore and Metcalf 1979). For Nan Madol, the fol- the social position of the deceased may be lowing ranking is proposed based on known evident in associated mortuary architecture. prevalence, labor investment, and visibility If so, patterning in burial architecture (see Ayres and Mauricio 1997:9–13): should directly relate to social stratification. Burial structures typically display relative political status rather than societal rank 1. Lolong burial complexes and bone (Kirch 1980:306). Sociopolitical status in life caches with the highest rank; should be directly related to social hierarchy 2. Burials in house platforms with middle in burial with larger structures indicating rank; higher status. Christian (n.d.:36) notes that 3. Other burials not associated with a spe- architectural scale (i.e., monumentality) can cific structure with the lowest rank. function to legitimate a social or political

446 VOLUME 6 • ISSUE 3 • 2011 Mortuary Contexts and Social Structure

group. In many instances architectural scale to the use of a structure or space. Non-burial can be related to time, where smaller struc- contexts would typically be expected to con- tures are older. The repeated use of lolong tain more utilitarian artifacts (e.g., stone and as mortuary monuments at Nan Madol added shell tools) than non-utilitarian artifacts or to the status associated with them during prestige items (e.g., beads, pendants and the history of their use. Over time successive jewelry), which would be more prevalent political leaders may have constructed larger in burial contexts. Variation in artifact fre- structures to legitimate their rule in relation quencies or type based on burial category to past leaders and contemporaneous rivals. might also suggest variation in social status Although larger burial structures are likely (Binford 1971; Saxe 1970; Tainter 1978). Ar- to reflect sociopolitical status, it is important tifacts associated with higher status burials to remember that change in architectural (i.e., lolong) should be distinct from those scale over time may not reflect status alone. non-burial contexts, since they should con- The lack of a chronology can obscure our tain some items that are not present in other ability to associate specific structures with contexts. Lower status burials are expected sociopolitical status based on construction tocontainamixofartifacttypesfoundinhigh size. status burial and non-burial assemblages. Spaces with formalized functions, like mortuary sites, typically have a spatial pattern (Clark and Martinsson-Wallin 2007:34). “The deliberation that lay behind the inter- Question 3: Is There Evidence for ment of the dead encodes social meaning Hierarchy in the Paleodemographic Burial in spatial order” (Ashmore and Geller Data at Nan Madol? 2005:91). Mortuary sites are often clustered or grouped together in specific areas. Such In societies with achieved status, rank groupings provide information on how a is expected to correlate with age (Binford society regarded death, and multiple group- 1971:21); demographic data from burial ings can suggest changes in preferred burial contexts would show age differences location over time. Mortuary sites are often in burial type and location, and artifact located on the periphery of a site as a way associations. Infants and young children are of separating the living from the deceased, generally treated differently in burial from though this is not always the case. If mor- adults in these societies, since they may not tuary locations are scattered across a site, it be considered full members of the society is plausible that the society did not see the (Hutchinson and Aragon 2002:45–46; Saxe need to formally separate the dead from the 1970:67). In societies where rank is ascribed living. one would expect greater variability in mortuary practices. This type of social structure may display differential treatment based on Question 2: Are Assemblages of Burial age and sex. Infants and children can be Artifacts Distinct From Non-Burial associated with the entire range of burial Contexts and Do Burial Artifact practices (Binford 1971; Saxe 1970:67; Assemblages Differ by Burial Type? Tainter 1978:106; Ucko 1969), which places the rank of some children over that of Artifact distributions may not reflect the some adults. In status-ascribed societies sex entirety of materials used at a site, but re- differences in burial demographics may be covered burial artifact assemblages should suggestive of either matrilineal or patrilineal be distinct from artifact assemblages found inheritance, or some other aspect of social in non-mortuary contexts at Nan Madol. Ar- structure (e.g., soldiers or priests being tifacts tend to fall into either a utilitarian separate from the general population). or non-utilitarian category. Artifact assem- Burial demographics at Nan Madol are blages from different contexts should show examined to see if they fit an achieved or patterning in artifact frequencies that relate ascribed pattern of status acquisition.

JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 447 Katherine Seikel

Question 4: Does Evidence From distribution of lolong is mapped separately Architecture, Artifact Distributions and to show that lolong are typically located on Paleodemographics Lead to Similar the seawall of the site or on the periphery Conclusions? of the site nearest Temwen Island. The distribution of other burial contexts is not spe- Each data-set should suggest the pres- cific to an area of Nan Madol and they are ence or absence of social hierarchy in mor- not highly visible, like lolong. This suggests tuary contexts at Nan Madol. Some types of that visibility potentially played a large role data may be more informative about social in mortuary treatment at Nan Madol. Individ- structure than others. For example, demo- uals with high sociopolitical status, the most graphic data can show evidence of differ- visible members of society, were buried in ences in burial based on sex or age, while a manner reflecting their status in life, in lo- artifactual data may inform on the presence long, while other, less visible members of or absence of social hierarchy. Supposed society would be buried in other contexts absence of social hierarchy from one data- (Huntington and Metcalf 1979; Saxe 1970). set does not mean social hierarchy was not Is there a link between the visibility of the lo- present within a society. The analyses of the longandtheirplacementontheseawall?Ash- data-sets should suggest similar patterning in more and Geller (2005:91) note that there sociopolitical status. If these data are consis- is social meaning in burial location. Placing tent, they can provide a framework to recon- lolong on the periphery of the site would struct social hierarchy at Nan Madol. have made these structures the first visible to individuals approaching the site, suggesting the power and authority of Nan Madol. In contrast, Binford (1971:12) noted the re- ARCHITECTURAL ANALYSIS lationship between environment and burial location. The seawall functions as a buffer This architectural analysis focuses on the between the sea and the rest of the site so monumental scale lolong tomb complexes. using the seawall islets primarily for burial The built environment constructs social re- may have been purely practical. lations (Paynter 1989:370), so architecture Measurements of area, volume and is particularly positioned to address social building orientation have been useful in structure in the archaeological record. Clark discussions of social structure and hierarchy and Martinsson-Wallin (2007:31) stated that in the Pacific and elsewhere (see Christian “monumental architecture is a structured n.d.; Clark et al. 2008; Graves 1986; Kirch venue for the production and transmission 1980, 2004). Architectural features were of ideas, traditions and belief systems.” mapped in plan view with cross-sections Researchers have noted that the ma- by Ayres (n.d.a.—Pahnwi A and Pahndipap) jority of Nan Madol’s tombs (lolong)and and Athens (n.d.—Karian and Pein Kitel). mortuary structures are located on the sea- These plan view maps were scanned and wall (Bath and Athens 1990:279; Hambruch then digitized in ArcGIS to determine areas 1936), but other mortuary contexts (bone and orientations for each lolong and their caches, house platform burials, etc.) are not associated platforms (Figure 4). Volume included in these determinations. Consider- calculations were taken for the structures ing the range of mortuary contexts, lolong based on measurements from map cross- are likely high status burial places since they sections. Lolong enclosures, platforms, and are among the largest structures at the site. orientations are all discussed separately. Distributions of islets with burial contexts Statistics were not used in this portion of the were mapped for Nan Madol to examine dif- analysis due to small sample size. ferences in burial location. Two maps were Five lolong complexes were present on produced to show the differences in distribu- the four architecturally mapped islets. These tion and frequency by islet based on the type complexes incorporated eight separate en- of mortuary context present (Figure 3). The closing walls. Lolong enclosure walls were

448 VOLUME 6 • ISSUE 3 • 2011 Mortuary Contexts and Social Structure

Figure 3. The distribution of burial contexts at Nan Madol: A.) Non-lolong burial contexts; B.) Lolong complexes. The numbers denote the number of context per islet.

JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 449 Katherine Seikel

Figure 4. Detail map of mortuary architecture: A.) Pahndipap; B.) Pein Kitel; C.) Pahnwi A; and D.) Karian. Note: The Karian and Pein Kitel architectural maps are copyrighted by Athens, and those of Pahndipap and Pahnwi A are copyrighted by Ayres.

compared based on area, volume, and the Pein Kitel, were ranked separately based on amount of space they enclose. There was a each individual measurement (Table 1). In wide range of variation based on overall en- each category the enclosure on Karian, the closure size (Table 1). Pein Kitel stands out southern enclosure on Pahnwi A, and the from the four other lolong complexes be- northern interior enclosure on Pein Kitel cause it has an enclosing wall surrounding ranked in the top three (wall areas above three separate lolong; so it is a single com- 150 m2), which suggests higher status. The plex with multiple enclosures. The added southern enclosure on Pein Kitel and the islet enclosure suggests that it had higher northern enclosure on Pahnwi A fall in the status than the other complexes based on in- middle (wall areas ∼90 m2). Pahndipap’s en- creased architectural complexity. The islet closure and the middle enclosure on Pein enclosing wall on Pein Kitel was excluded Kitel ranked at the bottom (wall areas ∼70 from the analysis because it served as an en- m2) suggesting lower status. There is a dis- closing wall for three lolong enclosures, and tinctive separation between the top three would be a signiﬁcant outlier based on func- enclosures and the rest of the sample, but tion and size. The other lolong enclosures, the distinction between middle and low sta- including the three smaller enclosures from tus is not particularly clear. The northern

450 VOLUME 6 • ISSUE 3 • 2011 Mortuary Contexts and Social Structure

Table 1. Measurements and overall rankings for each lolong enclosure.

Lolong enclosure walls Area (sq. m) Height (m) Volume (cu. m) Rank

Karian: Lolong enclosure 534.8 5.6 2,971.50 1 Pahndipap: Lolong enclosure 70.9 0.6 42.5 7 Pein Kitel: Surrounding enclosure 1,441.70 11.5 16,579.20 North enclosure 255.8 2.2 568.4 2 Middle enclosure 68.3 1 66.4 6 South enclosure 89.5 1.5 136.8 5 Pahnwi A: South enclosure 170.7 4.2 723.2 3 North enclosure 93 2.9 265.1 4

enclosure on Pein Kitel and the Karian enclo- is unclear and radiocarbon dates are needed sure both have two burial vaults, but the Pein to address this topic further. It is likely that Kitel enclosure only has one platform while additional architectural features, like the sur- Karian has two. The southern enclosure on rounding enclosing wall on Pein Kitel, in- Pahnwi A only has a single burial vault in a creasedassociatedstatusbyaddingcomplex- single platform, so the number of platforms ity to lolong complex. It would be useful in and burial vaults does not have a significant the future to compare the complexity in ar- impact on enclosure size. Karian is the only chitectural form on Pein Kitel to the islet of isletinthissamplewhereasinglelolongcom- Nahndauwas, which has a similar architec- plex encompasses the entire islet. Ethnohis- tural composition. torical evidence suggests that Karian func- Burial platform areas and volumes, like tioned as a priestly burial ground (Hambruch lolong enclosure size, may reflect social hier- 1936), which may be related to its large size archy. Eight burial platforms from the seven and architectural form. The northern enclo- lolong reported on above and one house sure on Pein Kitel has been associated with platform used for burial on Pahnwi A were Isohkelekel (Hambruch 1936), an important compared using area, height, and volume figure in Pohnpeian history, which may ex- (Table 2). All the platforms except the north- plain its significant size and its location in the ern Pahnwi A lolong platform and the Pah- inland portion of the site. The enclosure on ndipap platform fell within an area of 20– Pahndipap is the smallest in the sample, and 30 m2. Platform height ranges from 0.5–1.5 it has been suggested that Pahndipap is the m., which suggests that variation in platform oldest of the sampled islets (Ayres, personal height may be more important to studying communication 2007). In this case, Pahndi- hierarchy than area (Christian n.d.:30). Com- pap’s enclosure size may represent an ear- paring volume estimates shows three group- lier phase of lolong construction rather than ings in platform size: ∼10 m3, ∼20 m3,and reflecting differential status. Though there between 30 and 50 m3. The southern Kar- does seem to be a hierarchy in enclosure size ian platform, the Pahndipap platform and the based on individual lolong, it is difficult to house platform fall into the smallest group. know why lolong size may have changed This is expected because the southern Kar- over time. The relationship between time ian platform may be a secondary addition to and the perceived hierarchy of these lolong the lolong and the lolong on Pahndipap may

JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 451 Katherine Seikel

Table 2. Measurements and overall rankings for each burial platform.

Burial platforms Area (sq. m) Height (m) Volume (cu. m) Rank

Karian: North lolong platform 20.8 1.7 40.7 1 South lolong platform 20.4 0.6 11.3 7 Pahndipap: Lolong platform 13 0.9 11.7 8 Pein Kitel: North lolong platform 20.2 1 19.6 4 Middle lolong platform 21.1 0.8 17.6 6 South lolong platform 26 0.7 18.1 5 Pahnwi A: South lolong platform 28.7 1.5 42.3 2 North lolong platform 55.6 0.5 30 3 House platform 20.5 0.5 10.2 9 be older than the others, suggesting platform construction (i.e., Graves 1986; Kirch 2004). sizes increased over time. The house plat- Orientations for seven of the eight lolong form is expected to be small since it is not enclosing walls mapped by Ayres and Athens associatedwithanenclosureanddidnotorig- were established. The southern lolong on inally function as a burial site. The northern Pein Kitel was excluded because the original Karian platform is expected to be large since entrance could not be determined from the Karian was designated as the priestly burial map, which suggests that it is not preserved. ground (Hambruch 1936). The lolong plat- Long axis orientations corresponded with forms from Pahnwi A would seem to be the the facing orientations of the entrances. All outlierssincetherearenoadditionaldatasug- of the enclosures except the northern enclo- gesting a reason for the difference. The Pah- sure on Pahnwi A are oriented toward the nwi A platforms, being significant outliers, southwest. The northern enclosure on Pah- are suggestive of architectural hierarchy in nwi A is oriented to the northwest. None of platform size; this may be impacted by tem- the lolong are aligned with a cardinal direc- poral change and other factors such as re- tion, but this does not mean that structural modeling because the range of variation is so orientation lacks meaning (Kirch 2004:103). limited. Davidson (1967:91–92) noted that Unfortunately, published oral traditions and burial platforms elsewhere on Pohnpei are ethnohistorical accounts do not document similar in size to those found at Nan Madol, anything that notes that specific directions whichsuggeststhatoverallplatformsizemay hadculturalassociations,buttheremaybein- have been more standardized than enclosure formationthathasnotbeenpublishedonthis size. Testing the standardization of architec- subject. Further study, similar to Mauricio’s tural features with a larger sample from Nan (1987) work with clan narratives and settle- Madol and other areas on Pohnpei would be ment, would be valuable in understanding beneficial. the social importance of specific directions. “Specific burial location relative to landscape and settlement can carry addi- Are Differences in Mortuary Architecture tional social and cosmological significance” Related to Social Hierarchy? (Ashmore and Geller 2005:85). Studies of structural orientation are well suited for Distribution analyses definitely show a discussing social meaning in architectural visual hierarchy of mortuary contexts at Nan

452 VOLUME 6 • ISSUE 3 • 2011 Mortuary Contexts and Social Structure

Figure 5. MapofPahnwishowinglabeledfeatureswhereartifactsusedinthisanalysiswerecollected from.

Madolbuthierarchywithinindividualmortu- to the whole range of burial practices. Yet, ary contexts (i.e., lolong) is still unclear. The before one can address differences between lack of absolute dates in architectural analy- burial contexts a broader understanding of sis limits what can be determined about so- how burial contexts vary from other con- cial structure (Christian n.d.:19). Even with texts is necessary. In 1984 Ayres collected ar- a lack of temporal data, lolong enclosure size tifacts from surface scatters and sub-surface seems to be a more reliable indicator of hi- testing on Pahnwi A from six locations (Fig- erarchy than platform size. Orientation anal- ure 5). Feature 16, a non-burial context, was ysis paired with architectural size analyses the only area in which materials were ob- provides useful results for explaining archi- tained solely from surface collection. Ayres’ tectural hierarchy, but more data need to be artifact data from Pahnwi A were compared collectedrelatingtosocialmeaningsoforien- by context in order to determine variation in tation(i.e.,astronomicassociations).Further artifact assemblages based on depositional research related to architectural patterning location. Features were divided into cate- is necessary to clarify social structure at Nan gories: non-burial, non-structural burial, and Madol. structural burial (Table 3). These three categories were compared statistically using chi-square analyses and ARTIFACT ANALYSIS showed a signiﬁcant difference between the each context, whether shell beads were in- Saxe (1970:3) noted that the signiﬁcance of cluded (p < .001, χ 2 = 4959.37) in the one burial type is only apparent in relation sample or not (p = .001, χ 2 = 29.39).

JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 453 Katherine Seikel

Table 3. Pahnwi A artifact distributions by feature. Data is listed as percentages of total feature assemblage.

Artifact categories F16 F32 F1 F18 F23 F30 Total

Ceramics 0 0 0 0 2.70% 0 3 Stone tools and implements 28.60% 30% 43.30% 21.90% 11.60% <0.1% 44 Shell tools and implements 7.10% 10% 3.30% 18.80% 5.40% <0.1% 17 Fishing gear 0 0 0 0 0 0.10% 5 Jewelry 0 40% 3.30% 3.10% 34.80% 99.80% 10,504 Worked shell 64.30% 20% 50% 56.30% 45.50% 0.10% 103 Total artifacts 14 10 30 32 112 10,478 10,676

F16 and F32 are non-burial features, F1 and F18 are non-structural burials, and F23 and F30 are structural burials.

Overall the structural burials contained sig- Are Assemblages of Burial Artifacts nificantly larger amounts of prestige items Distinct From Non-Burial Contexts and (i.e., shell beads, conus armbands), while Do Burial Artifact Assemblages Differ by the non-burial contexts contained higher fre- Burial Type? quencies of debitage, pounders, and other more utilitarian goods. Predictably, the ar- The statistical analyses of depositional tifact frequencies from the non-structural contexts on Pahnwi A suggest social hier- burials contained a mix of items found archy based on the type of burial context. in the other two contexts. This suggests Lolong are placed at the top of this hierar- that there is a difference in the Pahnwi A chy, followed by house platform burials and artifact frequencies based on depositional finally non-structural burials. This ranking of location. burial contexts is consistent with both the To clarify the relationship between architectural data and the proposed ranking structural burial contexts a second set presented earlier. of chi-square analyses compared artifact frequencies by structural burial type (house platform vs. lolong). This analysis showed that there was a significant difference PALEODEMOGRAPHY between the house platform and lolong contexts (with beads: p < .001, χ 2 = 5780.31; Paleodemographic data for Nan Madol were without beads: p < .001, χ 2 = 70.21). collected from two reports on different col- The lolong contained significantly higher lections of skeletal material (Pietrusewsky frequencies of shell beads, conus armbands, and Douglas n.d.; Tasa n.d.). The skeletal and pearl shell items, whereas the house remains were collected from 11 of the 20 platform contained higher frequencies of identified burial islets by Ayres, Athens, and shell adzes, worked shell and other more Bath in 1984 (Figure 6). The mortuary con- utilitarian items. Since non-structural burials texts include mortuary structures, lolong, contained few prestige items compared to platforms, bone caches, and non-structural the lolong and platform burials, this suggests burials. The prevalence of secondary bone that house platform burials rank between lo- deposits, implying secondary burial prac- long and non-structural burials in the burial tices, was high across the burial types. While hierarchy. some ‘mortuary structures’ may be lolong,

454 VOLUME 6 • ISSUE 3 • 2011 Mortuary Contexts and Social Structure

Figure 6. The distribution of mortuary islets. The numbers denote the number of individual identi- ﬁed from the islets analyzed in this study.

Pietrusewsky and Douglas’s report (n.d.) is teristics, and cranial morphology but many not clear on this matter so they are treated of individuals could not be sexed due to the separately. The following results should be poor condition of the remains. Comparisons considered preliminary. were made between burial types, age groups A total of 51 individuals were identified and sex (Table 5). by sex and age, where possible (Tables 4 In egalitarian societies the oldest mem- and 5). Both reports comment on the poor bers of a group tend to have the highest sta- preservationoftheskeletalmaterial,whichis tus, which would be reflected in the burial generally fragmentary and badly weathered context (Binford 1971:21); whereas, infants (Pietrusewsky and Douglas n.d.:1). Identifi- and children are typically buried separate cations of age and sex were based on robus- from adults, since they would not acquire ticity, size, muscle markings, tooth charac- group status until they approach adulthood

JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 455 Katherine Seikel

Table 4. Paleodemographic breakdown by burial type and islet shown as counts.

Burial type Infant (0–1) Child (1–12) Adolescent (12–20) Adult (20 + ) Total

Lolong (n = 1) 8 Pahnwi A 0 2 0 6 Mortuary structure (n = 5) 22 Lemenkau 0 0 0 3 Likinpei A 1 2 0 2 Pahndauwas 0 0 0 3 Peinering 0 0 0 2 Peinioar 0 2 0 7 Platform (n = 5) 13 Pahndauwas 0 0 0 5 Pahnwi A 0 1 0 2 Sapwenpwe 0 1 0 2 Sapwuhdir 0 1 0 0 Wasau 0 0 1 0 Bone cache (n = 5) 6 Nahndauwas 0 0 0 1 Pahnwi A 0 2 0 0 Pohndauwas (n = 3) 0 0 1 2 Other burial (n = 2) 2 Pahnwi A 0 0 0 2

The classiﬁcation of mortuary structure may overlap with other classiﬁcations. Pahnwi A and Wasau data taken from Tasa n.d. and all other data from Pietrusewsky and Douglas n.d.

(Saxe 1970:67). Therefore, comparing burial levels. For this reason, subadult remains are contexts based on age is important in un- less likely to survive in the taphonomic en- derstanding social structure. Subadults com- vironment of sub-surface interments. Statis- posed between 20 and 30 percent of the in- tical significance of burial location based on terred population in lolong and platforms agewasdeterminedusingachi-square.There and 50 percent for bone caches, but none was no significant difference ( p = .621, χ 2 = were associated with non-structural burials. 2.63) in individuals being buried in locations Subadults are defined as individuals under based on age. Since subadults are present in the age of 20 (Ubelaker 1999:63). This vari- nearly all burial contexts and age estimates ability in age distribution may be due to for the subadult population range between taphonomicbiasbasedonburiallocationand 18 months and 15 years (Pietrusewsky and type rather than differential treatment of age Douglas n.d.; Tasa n.d.), the data suggest groups. Non-structural burials are typically that social status was inherited rather than interments in the islet surface. Taphonomi- achieved. The presence of subadults in high cally these areas are constructed of coral rub- status burial contexts supports this. ble and waters leach into the lower portion Overall more adult female skeletal re- of the islets causing coral to breakdown over mains were identified than male. Based on time, so small bones or fragments may be lost raw counts there seem to be more females in gaps created between coral rubble pieces associated with mortuary structures than that shifted due to coral breakdown in lower males, but with the small sample size and

456 VOLUME 6 • ISSUE 3 • 2011 Mortuary Contexts and Social Structure

Table 5. Paleodemographic breakdown by burial type shown as counts and percentages. The table also records the number of sampled contexts and the total for each burial type.

Adult unidenti- Adult Adult Burial type Subadult Adult ﬁed male female Total

Lolong (n = 1) 2 (25%) 6 (75%) 6 (75%) 0 0 8 Mortuary structure 5 (22.7%) 17 (77.3%) 5 (22.7%) 3 (13.6%) 9 (40.9%) 22 (n = 5) Platform (n = 5) 4 (30.8%) 9 (69.2%) 6 (46.2%) 1 (7.7%) 2 (15.4%) 13 Bone cache (n = 5) 3 (50%) 3 (50%) 0 1 (16.7%) 2 (33.3%) 6 Other burial (n = 2) 0 2 (100%) 1 (50%) 1 (50%) 0 2

The classification of mortuary structure may overlap with other classifications. the large number of adults of indeterminate CONCLUSIONS sex this relationship is unclear. A second chi- square, excluding indeterminate individuals, Does Evidence From Architecture, Artifact determined that there was no significant dif- Distributions and Paleodemographics ference (p = .660, χ 2 = 2.42) in burial loca- Lead to Similar Conclusions? tionbasedonsex.Inahereditaryhierarchical system it is possible for sex differences to be The architectural and artifactual data present in a mortuary context, but that does show that, as expected, social hierarchy is not seem to be the case at Nan Madol based present in the archaeological context of Nan on this data-set. Madol. There are differences in artifacts associated with different burial contexts, and variation in architectural distribution and Is There Evidence for Hierarchy in the size suggestive of hierarchical differences. Paleodemographic Burial Data at Nan The apparent social hierarchy in these data- Madol? sets is supported by the identification of ascribed social status in the paleodemographic Based on presence of subadults in lo- data. Taken together the archaeological data long, mortuary structures, platforms, and identify a hierarchical social system that is bone caches socio-political status at Nan based on inherited socio-political status at Madol was likely ascribed at birth. There was the site of Nan Madol. no statistical difference in burial based on The archaeological data presented in sex, so differential burial or status based on this study parallels information from Pohn- sex is not likely at Nan Madol based on this peian oral traditions, which state that Pohn- data-set. Unfortunately, these are only pre- peihadahierarchicalchiefdomsystembased liminary conclusions due to the small sample on ascribed status within clan units exist- sizes and the lack of chronological data as- ing during Nan Madol’s time as a political sociated with these mortuary contexts. The center (Bernart 1977; Hambruch 1936; Han- high prevalence of secondary burials may lon 1988). This is not the first exam- have also impacted the sample but was not ple of an archaeological study which sup- able to be more fully addressed due to the ports information from oral traditions and paucity of information on secondary burial ethnohistory on Pohnpei (Athens 2007; practices on Pohnpei. Further studies will be Ayres 1990; Bath and Athens 1990), but necessary to clarify questions on chronology it is the first to utilize data from mortu- and burial practice. ary architecture, paleodemographics, and

JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 457 Katherine Seikel

artifacts to discuss social hierarchy at Nan thanks go to my friends and colleagues at Madol. ANUandUniversityofOregonfortheirhelp This study represents a starting point for andsupportinvariousphasesofthisunder- future research on mortuary practices and taking. social hierarchy on Pohnpei. Patterns in archaeological and oral traditional data are useful in understanding mortuary practices in the context of Pohnpeian history, but fur- REFERENCES ther research is required to clarify this relationship. Current information on Pohnpeian Anton, S. C. and D. W. Steadman. 2003. Mortu- burial practices is primarily limited to identi- ary patterns in burial caves on Mangaia, Cook fied burial types. Little is known about how Islands.International Journal of Osteoarchae- mortuarycontextsrelatetoeachotheracross ology 13:132–146. Pohnpei’s history and obtaining dates from Ashmore, W. and P. Geller. 2005. Social dimen- mortuary contexts will clarify the relations sions of mortuary space. In Interacting with among mortuary patterning, time, and social the Dead: Perspectives on Mortuary Archae- structure. ology for the New Millennium (G. Rakita, J. “Chiefdoms are probably best defined Buikstra, L. Beck, and S. Williams, eds.):81– 92. Gainesville, FL: University of Florida as regionally organized societies with a cen- Press. tralized decision-making hierarchy coordi- Athens, J.S. 1980. Archaeological Investigations nating activities among several village com- at Nan Madol: Islet Maps and Surface Arti- munities” (Earle 1987:288). Based on this facts. Agana: Pacific Studies Institute. definition, the presence of the monumen- Athens, J. S. 1984. Surface artifact distributions at tal constructions of Nan Madol and Sapw- the Nan Madol Site: A preliminary assessment takai in southern Pohnpei supports the iden- of spatial patterning. New Zealand Journal of tification of Pohnpeian society as a system Archaeology 6:129–153. of chiefdoms. Nan Madol is physically sepa- Athens, J. S. 1990. Nan Madol pottery, Pohnpei. rate from the rest of Pohnpeian society (Bath Micronesica Supplement 2:17–32. Athens, J. S. 2007. The rise of the Saudeleur: Dat- and Athens 1990:280), so there may be dis- ingtheNanMadolchiefdom,Pohnpei.InVastly tinct differences in mortuary and settlement Ingenious: The Archaeology of Pacific Mate- patterning based on this separation. Further rial Culture (A. Anderson, K. Green, and F. archaeological research on Pohnpei is nec- Leach, eds.):191–208. Dunedin: Otago Univer- essary to clarify the relationship between sity Press. chiefly centers. Intersite comparisons will Athens, J. S. n.d. Completion Report for Develop- be integral to understanding social complex- ment of the Nan Madol Ruins, Ponape State, ity across Pohnpei over its entire history. FSM. Submitted to the Trust Territories Historic Temporal patterning of archaeological data Preservation Office, 1984. from across the island is necessary to under- Ayres, W. S. 1990. Pohnpei’s position in Eastern Micronesian Prehistory. Micronesica Supple- standing Pohnpei’s past and future research ment 2:187–212. should address the entire history of Pohnpei Ayres, W. S. 2002. Island archaeology and issues of and its position amongst the other complex political centralization: Micronesian evidence. societies of the Pacific. In World Islands in Prehistory: International Insular Investigations (W.H.WaldrenandJ.A. Ensenyat, eds.):57–67. BAR International Series ACKNOWLEDGEMENTS 1095. Oxford: Archaeopress. Ayres, W. S. n.d.a. Base Map of Nan Madol and Maps of Pahndipap and Pahnwi.Onfileat I thank William Ayres, Geoff Clark, Jon Er- the University of Oregon (used courtesy of W. landson, and the reviewers for their help- S. Ayres). ful comments and critiques. I would also Ayres, W. S. n.d.b. The Archaeology of Nan like to recognize the researchers who made Madol, Ponape, Micronesia (Eastern Caroline their unpublished reports and raw data Islands). Report submitted to the National Ge- available for this study. And finally, many ographic Society, Washington, DC, 1985.

458 VOLUME 6 • ISSUE 3 • 2011 Mortuary Contexts and Social Structure

Ayres, W. S. n.d.c. Nan Madol Archaeological Fischer J., S. Riesenberg, and M. Whiting. 1977. Fieldwork. Final Report submitted to Pohnpei Annotations to the Book of Luelen. Honolulu: State Historic Preservation Office, FSM, 1993. University of Hawaii Press. Ayres, W. S., A. Haun, and C. Severance. 1981. Fitzpatrick, S. and G. Nelson. 2008. From lime- Ponape Archaeological Survey: 1978 Re- stone caves to concrete graves: 3000 years of search. Micronesian Archaeological Survey Re- mortuary practice in the Palauan Archipelago, port 4. Saipan: Micronesian Archaeological Sur- Micronesia. International Journal of Osteoar- vey. chaeology 18:439–457. Ayres, W. S. and R. Mauricio. 1997. Salapwuk Graves, M. 1986. Organization and differentiation Archaeology: A Survey of Historic and Cul- within Late Prehistoric ranked social units, Mar- tural Resources on Pohnpei, Federated States iana Islands, Western Pacific. Journal of Field of Micronesia. San Francisco: Micronesian En- Archaeology 13(2):139–154. dowment for Historic Preservation and the U.S. Gulick, L. H. 1859. Micronesia: The ruins on National Parks Service. Ponape, or Ascension Island. Journal of the Ayres, W. S. and C. Scheller. 2002. Status archi- American Geographical and Statistical Soci- tecture and stone resources on Pohnpei, Mi- ety 1(5):129–137. cronesia: Experiments in stone transport. In Hambruch, P. 1936. Results of the South Seas Ex- Fifty Years in the Field. Essays in Honour and pedition: 1908–1910 (Vol. 7, Ponape Part 3) Celebration of Richard Shutler Jr’s Archaeo- (Murphy and Runeborg, trans.). On file at Uni- logical Career (S.Bedford,C.Sand,andD.Bur- versity of Hawaii. ley, eds.):109–121. Auckland: New Zealand Ar- Hanlon, D. 1988. Upon a Stone Altar: A History chaeological Association. of the Island of Pohnpei to 1890. Honolulu: Bath, J. and J. S. Athens. 1990. Prehistoric social University of Hawaii Press. complexity on Pohnpei, the Saudeleur to Nah- Hughes, D.T. 1982. Continuity of indigenous Pon- nmwarki transformation. Micronesica Supple- apean social structure and stratification. Ocea- ment 2:275–290. nia 53(1):5–18. Bernart, L. 1977. Book of Luelen ( J.Fischer,S. Huntington, R. and P. Metcalf. 1979. Celebra- Riesenberg, and M. Whiting, eds. and trans.). tions of Death: The Anthropology of Mortu- Pacific History Series No. 8. Canberra: Aus- ary Ritual. New York: Cambridge University tralian National University Press. Press. Binford, L. 1971. Mortuary practices: Their study Hutchinson, D. and L. Aragon. 2002. Collective and their potential. In Approaches to the Social burials and community memories: Interpreting Dimensions of Mortuary Practices (J. Brown, the placement of the dead in the southeastern ed.):6–29. Memoirs of the Society for American United States with reference to ethnographic Archaeology. cases from Indonesia. In The Space and Place Christian, T. n.d. Monumentality and function- of Death (Silverman and Small, eds.):27–54. Ar- ality: Some dimensions of and approaches to chaeological Papers of the American Anthropo- the analysis of monumental architecture for ar- logical Association 11. chaeology. Unpublished Master’s Paper, Uni- Kirch, P. 1980. Burial structures and societal rank- versity of Oregon, 1998. ing in Vava’u, Tonga. Journal of the Polynesian Clark, G., D. Burley, and T. Murray. 2008. Society 89:291–308. Monumentality and the development of the Kirch, P. 1990. Monumental architecture and Tongan maritime chiefdom. Antiquity 82: power in Polynesian chiefdoms: A compari- 994–1008. son of Tonga and Hawaii. World Archaeology Clark, G. and H. Martinsson-Wallin. 2007. Mon- 22(2):206–222. umental architecture in West Polynesia: Ori- Kirch, P. 2004. Temple Site in Kahikinui, Maui, gins, Chiefs and Archaeological Approaches. Hawaiian Islands: Their orientations decoded. Archaeology in Oceania Supplement 43:28– Antiquity 78(299):102–114. 40. Kirch, P. and J. L. Rallu (eds.). 2008. The Growth Cordy, R. 1985. Settlement patterns of complex and Collapse of Pacific Island Societies: Ar- societies in the Pacific. New Zealand Journal chaeological and Demographic Perspectives. of Archaeology 7:159–182. Honolulu: University of Hawaii Press. Davidson, J. 1967. Preliminary archaeological in- Lilley, I. (ed.). 2006. Archaeology of Oceania: vestigations on Ponape and other Eastern Caro- Australia and the Pacific Islands. Oxford: line Islands. Micronesica 3:81–95. Blackwell Publishing. Earle, T. 1987. Chiefdoms in archaeological and Liston, J. 2009. Cultural chronology of earthworks ethnohistorical perspective. Annual Review of in Palau, Western Micronesia. Archaeology in Anthropology 16:279–308. Oceania 44(20):56–73.

JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 459 Katherine Seikel

Mauricio, R. 1987. Peopling of Pohnpei Island: Scott,R.,H.Buckley,M.Spriggs,F.Valentin,and Migration, dispersal and settlement themes in S. Bedford. 2010. Identification of the first re- clan narratives. Man and Culture in Oceania ported Lapita cremation in the Pacific Islands 3:47–72. using archaeological, forensic and contempo- Parke, A. 1998. Navatanitawake ceremonial rary burning evidence. Journal of Archaeolog- mound, Bau, Fiji: Some results of 1970 inves- ical Science 37(5):901–909. tigations. Archaeology in Oceania 33:20–27. Silverman, H. and D. Small (eds.). 2002. The Space Paynter, R. 1989. The archaeology of equality and and Place of Death. Archaeological Papers of inequality. Annual Review of Anthropology the American Anthropological Association 11. 18:369–399. Tainter, J. 1978. Mortuary practices and the study Peebles, C. and S. Kus. 1977. Some archaeological ofprehistoricsocialsystems.InAdvances in Ar- correlate of ranked societies. American Antiq- chaeological Method and Theory 1 (M. Schif- uity 42(3):421–448. fer, ed.):105–141. New York: Academic Press. Pietrusewsky, M. and M. Douglas. n.d. Report Tasa, G. n.d. Report on human skeletal remains on Human Skeletal Material Recovered from from Pahnwi and Wasau, Nan Madol, Pohnpei. Nan Madol Islets, Ponape. Report completed Unpublished Master’s Paper, University of Ore- 1985, University of Hawaii. gon, 1988. Pietrusewskey, M., M. Douglas, E. Cochrane, and Ubelaker, D. 1999. Human Skeletal Remains: S. Reinke. 2007. Cultural modifications in an Excavation, Analysis, Interpretation,3rded. adolescent earth-oven interment from Fiji: Sort- Washington, DC: Taraxacum. ing out mortuary practice. Journal of Island Ucko, P. 1969. Ethnography and archaeological and Coastal Archaeology 2(1):44–71. interpretation of funerary remains. World Ar- Rainbird, P. 2004. The Archaeology of Microne- chaeology 1(2):262–280. sia. Cambridge: Cambridge University Press. Valentin, F. 2003. Human skeletal remains from Rakita, G., J. Buikstra, and L. Beck (eds.). 2005. In- the site of Lapita at Kone (New Caledonia): Mor- teracting with the Dead: Perspectives on Mor- tuary and biological features. In Pacific Archae- tuary Archaeology for the New Millennium. ology: Assessments and prospects (C. Sand, Gainesville, FL: University of Florida Press. ed.):285–293. Noumea: Service des Musees´ et Saxe, A. 1970. Social Dimensions of Mortuary du Patrimoine de Nouvelle-Caledonie.´ Practices. Dissertation, University of Michigan. Valentin, F., S. Bedford, H. Buckley, and M. Saxe, A., R. Allenson, and S. Loughridge. 1980. Spriggs. 2010. Lapita burial practices: Evidence The Nan Madol Area of Ponape: Researches for complex body and bone treatment at the into Bounding and Stabilizing an Ancient Ad- TeoumaCemetery,Vanuatu,SouthwestPacific. ministrative Center. Saipan: Historic Preven- Journal of Island and Coastal Archaeology tion Office, Trust Territory of the Pacific. 5(2):212–235.

460 VOLUME 6 • ISSUE 3 • 2011 Copyright of Journal of Island & Coastal Archaeology is the property of Routledge and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.