25. From Thatch to Cement:
Social Implications of Housing Change on Rotuma
[Published inHome in the Islands: Housing and Social Change in the Pacific, edited by Jan Rensel and Margaret Rodman. Honolulu: University of Hawaii Press, 1997]
For centuries, houses--their construction, maintenance, use, and even their location--have been central to the social reproduction of kin groups (kainaga) on the island of Rotuma.  Kainaga membership is a matter of both blood relationship and active demonstrations of commitment. Blood ties are reckoned on the basis of a common ancestor who lived on or had claim in a named house site. And commitment to a kin group is demonstrated by giving materials and labor to building, maintaining and furnishing a house, as well as by being a part of activities that take place in and around it. By attending gatherings, contributing resources, helping prepare food, and eating together with those who dwell in a given house, Rotumans repeatedly proclaim their connectedness. Rotuman houses stand as tangible reminders and powerful symbols, embodying the responsibilities and relationships of all who participate in their construction, repair, and use.
Rotuma, like other Pacific Islands, has undergone significant social, economic and demographic change, especially during the past few decades. Studies elsewhere have pointed to the far-reaching effects of outmigration and remittances, for instance (see, e.g., Bertram and Watters 1985, 1986; Hooper and Huntsman 1973; Severance 1976; Shankman 1976, 1992; O'Meara 1986). Economic problems associated with these processes include declines in agricultural productivity, weakened potential for development, maintenance of high living standards by external subsidies, and consequent vulnerability to external economic fluctuations. Social impacts range from the erosion of traditional authority patterns and status structures, and incipient class formation based on material wealth, to the devaluation of collective work, community fragmentation, the spread of individualism, jealousy, and dissatisfaction. Many of the effects of population movement, shifting economic bases, and increased access to cash and imported materials are represented tangibly in Pacific Islands houses. Changes in house materials, construction, and use can provide a focal point for examining wider patterns of transformation.
Since the earliest recorded descriptions of Rotuman houses in the nineteenth century, there have been many documented changes in house styles, materials, and construction processes. Contributing factors range from missionary influence and increasing external trade, to hurricanes, relief programs, and migrant involvement. Changes in structures, and processes surrounding their construction, maintenance, and use, are affecting social relationships in some important ways. This chapter is concerned with tracing changes in housing on Rotuma, identifying significant factors producing those changes, and exploring the implications for Rotuman social relationships.
The island of Rotuma is relatively remote, located 465 kilometers north of the northernmost island in the Fiji group, and only slightly closer to Futuna, its nearest neighbor (see map; figure 2.1). Rotuma has been politically affiliated with Fiji for more than a century, first as a British colony and since 1970 as part of the independent nation. Rotuma's people are, however, culturally and linguistically distinct, having strong ties with Tonga, Samoa, and other Polynesian islands to the east.
The island is composed of seven districts, each of which has its own paramount chief and a number of subchiefs. A chief is selected from and by a special group of kin called a mosega (literally, 'bed'), who claim descent from a particular house site with which the chiefly title is associated. Although Rotumans cooperate on communal projects under the direction of their subchiefs and chiefs, and frequently engage in interhousehold exchange of food and labor, households are largely self-sufficient. Rotuma is a fertile volcanic island of forty-three square kilometers, surrounded by a fringing coral reef of varying width and productivity. The land supports the cultivation of a range of starchy staple crops as well as other vegetables, fruits, and coconuts. Most Rotuman households keep chickens and pigs, and some raise a few goats or cows as well. Local protein sources include meat from these animals, and fish, shellfish, and seaweed from surrounding waters.
In addition to subsistence production Rotumans have engaged in external trade for at least two centuries. Locally produced food formed the basis for commerce with whalers and other European ships from the time of first recorded contact with Europeans (HMS Pandora) in 1791. In addition, Rotumans eagerly signed aboard passing ships as crew and earned money as sailors and pearl divers. Trading in copra was established by the 1870s, and copra has remained Rotuma's chief export, despite fluctuations in production and declining profits in recent decades. After incorporation as a British colony in 1881, Rotuma was closed as a port of entry, so most opportunities for trade, employment, and education were pursued in and through Fiji. Although the population on Rotuma has remained relatively stable, the proportion of Rotumans away from the home island has grown steadily, such that in 1986, 70 percent of Rotumans were living in Fiji rather than in Rotuma (see table 2.1).
Rotumans on Rotuma and In Fiji, 1956-1986
Source: Fiji Census Reports, Government Press, Suva, Fiji
Employment opportunities on Rotuma more than doubled from 1960 to 1989. According to an island-wide survey conducted in 1960 by Alan Howard, 71 Rotumans held wage positions; a similar survey in 1989 recorded 174 Rotumans earning wages. The primary employers were the Fiji government (e.g., school teachers, medical staff, laborers) and two Rotuman cooperatives that handled copra sales and imported food and other supplies.  Of 414 households on Rotuma surveyed in 1989 (85 percent of all households on the island), 167 households (40 percent of those surveyed) included wage, pension, or self-employed earners. Household members also earn money from a variety of sources, including casual labor, sporadic exports of crops and animals, and on-island trade-for instance, selling garden produce to government workers. In addition, nearly half (48 percent) of the households surveyed reported receiving cash remittances from relatives in Fiji or abroad. Many households benefit as well from other forms of continuing involvement of Rotuman migrants, such as help obtaining material goods from Fiji, and assistance with small entrepreneurial projects, including handicraft sales and tourist visits (see Rensel 1993 regarding the importance of migrant involvement to the material well-being of Rotuma).
A mounting reliance on imported rather than local food is suggested in records of Rotuman chiefs, who monitor garden production in their respective districts. Between 1966 and 1981, production of most staple starch crops fell, in some cases dramatically. Crop plant counts for that fifteen-year period show, for instance, that taro (Colocasia) declined from 326,000 to 289,000 plants; yams (Discorea) from 101,000 to 69,000; and cassava (Manihot) from 332,000 to only 99,000. These drops cannot be attributed to a shortage of manpower. Not only has the island population remained fairly constant from 1956-1986, but the numbers of Rotuman men between the ages of fifteen and fifty, who do most of the gardening, have also held steady at around 500-600 (approximately 20 percent of Rotuma's population).
More revealing of possible causes for drops in local food production are the records of the Rotuma Cooperative Association (RCA), which handled most of the island's trade from the late 1950s until recently. RCA turnover figures from 1957 to 1986 clearly display a "jaws effect" as purchases of imported goods diverge from copra sales (see Bertram and Watters 1985, 510). Whereas consumer spending initially was closely tied to copra income, by 1986 RCA store purchases (F$1,022,790) outstripped copra sales (F$323,120) more than three to one.  Growing income from wages, remittances, and sources other than copra has served to fuel import consumption.
Although households use much of their income to purchase tinned or packaged foods, a more obvious result of increasing cash affluence can be found in changes in housing on the island. Rotumans have a long history of seeking to improve housing and to make construction and maintenance processes less labor intensive. In the following section I begin with some of the earliest available descriptions of Rotuman houses, before turning to the changes recorded over ensuing decades.
Building Materials and Styles
An idea of housing styles on Rotuma from the early 1800s can be gleaned from the accounts of some of the first European visitors (Bennett 1831; Eagleston 1832; Osborn 1834-1835; Cheever 1834-1835; Lesson 1838-1839; Lucatt 1851; Haley  1948). Houses were constructed of poles and logs, with thatched sago palm roofs and plaited sago or coconut palm walls. Most dwellings were described as "small," enclosing a space perhaps 15 to 20 feet wide (Eagleston 1832, 409). Chiefs' houses were noted as being larger, for instance 40 by 16 feet (Haley 1948, 259) and 25 feet high (Lesson 1838-1839, 433). These early written accounts describe Rotuman houses as rounded at the ends (see figure 2.2), but according to Elisapeti Inia, a retired Rotuman schoolteacher and recognized authority on Rotuman custom,  the rounding was due to Samoan or Tongan influence; the ends of Rotuman houses were originally flat (tarut fari).
2.2 Thatched Rotuman meeting house showing many elements common to traditional dwellings, though the plaited walls of the latter would be full rather than half height. The rounded ends of the thatched roof, recorded by early nineteenth-century observers, are attributed by Rotumans to Samoan and Tongan influence. Alan Howard, 1960.
Low doors, which admitted little wind as a protection against hurricanes, required people to enter on hands and knees. Floors were composed of earth, dry grass, and pebbles or small pieces of coral, covered with rough mats of plaited coconut leaves (farao); sometimes with a pandanus mat ('epa) overlay.
Cooking and eating took place outside or in a separate outbuilding (kohea 'kitchen'), also made of poles and thatch. In practice, these spaces were more accessible to members of other households than were the interiors of dwellings. This facilitated the Rotuman custom of assisting others with cooking in koua 'earth ovens' as well as that of sharing meals.
Young unmarried men ordinarily slept away from their parents and siblings. It was considered improper for them to sleep inside the house, in close proximity to their sisters. Groups of young men sometimes built their own thatched sleeping houses, sometimes on high poles (ri sipakit). By staying together, as well as by participating in other joint activities such as preparing koua and gardening, youths strengthened not only their relationships with each other but ties between their respective households.
Rotumans customarily built their houses on a foundation, or fuag ri, of raised earth, surrounded by stone walls (Osborn 1834-1835; Cheever 1834-1835; Lucatt 1851:167). Most reports indicate that foundations were from two to four feet high, but descriptions range from one foot (Allardyce 1885-1886, 134) to six feet high (Allen 1895). Foundations up to twelve feet high, presumed to have been used for chiefly dwellings, were discovered inland by Gardiner (1898, 433). Some writers suggested these raised house sites were useful in keeping the floors dry during periods of heavy rains (Osborn 1834-1835; Lucatt 1851, 167; Boddam-Whetham 1876, 266). For Rotumans, however, fuag ri were and are significant in notions of kinship. It is the house foundation to which Rotumans generally refer when they describe how they are related to someone, for example: "My mother's mother is from the fuag ri where he stays," or "He is related to that fuag ri." Fuag ri are also reference points for eligibility to stewardship of associated kainaga garden lands, and some foundations carry with them chiefly titles.
Home Furnishings and Housekeeping
Early visitors to Rotuma reported but little in the way of house furnishings: "mats, carved bare wood pillows, a few clubs, spears and drinking vessels of coconut shells" (Osborn 1834-1835). Lesson (1838-1839, 434) mentioned low tables for eating. Coconut shells strung on sinnet for carrying water could be hung up in the house (Eagleston 1832), and "in the centre of the house is generally slung a little koop net on which are deposited their provisions etc." (Cheever 1835). A more elaborate description of a storage device is given by W. L. Allardyce, who was acting resident commissioner in 1881:
There is scarcely a house which does not possess, suspended from the ridgepole, a kind of large four-sided swinging basket, called kokona, which serves as a larder and cupboard, and general receptacle for things which are intended to be out of the way of the children and rats. To guard against the latter a piece of circular wood, a foot or more in diameter, is obtained, and a hole bored in the centre, through which the main string of the kokona passes. Underneath this piece of wood, when a suitable height, a knot is made, not large enough to pass through the hole in the wood, which is thus kept stationary. However, the slightest weight on any part of it, at once gives the wood a sudden tilt downwards, and the rat is dropped on to the floor, clear of the kokona, and alongside of the cat. (Allardyce 1885-1886, 134)
Given the importance of mats as primary furnishings as well as items in ceremonial exchange, one could assume that plaiting them took up much of women's time. Mat making is often a cooperative activity, with women helping each other process pandanus and taking turns working on each other's mats. Cleaning a Rotuman-style house includes sweeping the floor, sunning the mats, and picking up leaves and other rubbish in the compound. Although some nineteenth-century European visitors found Rotuman houses "small, dark and dirty" (Forbes 1875, 227), others were impressed with how neat and "scrupulously clean" they were (see. e.g., Lesson 1838-1839, 434; Bennett 1831, 201; Haley 1948, 258). Timing may have affected observers' impressions: For instance, according to Rotuman custom, when men go out deep-sea fishing, women are not to clean the house. Similarly, for five days after a burial, houses of families in mourning remain unswept.
House Construction and Repairs
Customarily, Rotuman house building is a group process, although it may be guided by one who is particularly skilled (majau 'expert, carpenter'). Members of the kainaga assist, along with neighbors and friends. With thatched structures, women as well as men contribute materials and labor, helping to collect and prepare the poles and sago or coconut palm fronds. Host household members may also work on the building, but more of their efforts go toward providing food for the other workers. As noted in 1913 by A. M. Hocart, the host household prepares a feast for the majau both before and after the house is built, and provides meals for the workers every day on which they work (Hocart 1913, field note 4846). In addition, members of the household remain indebted to those who help them. They should be ready to reciprocate with their labor when needed. In his 1940 autobiography, Rotuman Methodist minister Rev. Fuata Taito describes the process:
If I had a big job to do, which would take a long time to do by myself, I would announce it in our village meeting. The chief always asks the question, "Has anyone a job to do which needs help?" Then he calls our names individually, and when he comes to my name, I reply, "Yes, sir, I want to build a hut, and would like ten men to help me if possible." He then appoints a suitable date for it, and calls on whoever can to put their own jobs aside for the day, and go and help Fuata to build his hut.
In addition to expectations of future reciprocal assistance, relatives who help with building projects may reap other benefits. A house on a fuag ri, or on kainaga land, is subject to use rights by members of the kainaga, and these claims can be strengthened by contributing labor toward construction.
Thatched buildings need periodic upkeep, providing further opportunities for demonstrating kin commitments. Rotumans valued sago palm as more durable than coconut palm for roofing thatch (Bennett 1831, 201; Evans 1951, note 25). According to the report of a Methodist minister who stayed on the island for several years in the 1880s, a sago palm roof "put on nicely is said to last without rethatching for twelve or sixteen years" (Allen 1895). To protect thatched roofs during strong winds, pairs of coconut palm fronds were (and are) laid over the roof vertically, tied together at the top. Still, thatched roofs and plaited walls must be periodically replaced. As with the process of building a new thatched structure, rethatching is an activity that typically involves a group of relatives and neighbors contributing materials and labor on a reciprocal basis and being thanked with food.
In their location on named fuag ri 'foundations,' and in the processes of their construction and maintenance, Rotuman-style thatched houses served as constant reminders to their inhabitants of the network of kin relations that supported them. Over the past century, however, several influences combined to effect wide-reaching changes in house materials, styles, construction and repair practices. In the next section I consider how a range of social, environmental, demographic, and economic variables have affected Rotuman housing standards, and in turn, the place of houses in the enactment of social relations.
Christian missionaries, who arrived on Rotuma in the mid-nineteenth century, affected housing on Rotuma both intentionally, as an explicit agenda, and indirectly, by introducing new building materials and techniques. The British Methodists in particular associated material lifestyle with spiritual orientation, and consciously tried to provide models of dress, cleanliness, and housing for Rotumans to emulate. Brother Osborne, writing from Sydney after leaving Rotuma in 1873, praised the work of his predecessors, Rev. and Mrs. William Fletcher and other Methodist teachers, and credited changes in housing on Rotuma to their efforts:
Before Wm. Fletcher's last appointment to the island, there was a comparatively large number of Christians, but they were necessarily very ignorant . . . their houses were the meanest hovels imaginable, and they themselves were unutterably filthy. . . . Through the instrumentality of Mr. & Mrs. Fletcher, and several really superior Fijian teachers, the most gratifying changes were effected. Hundreds lotu'd [entered the church] . . . then they purchased soap . . . then they grew dissatisfied with their hovels, and commenced the erection of substantial and neat houses. So rapidly did they advance, that when I was appointed to take Mr. Fletcher's place, nearly four years ago, I found that there was a membership of upwards of 450, & a large attendance at the schools. There were also scores of well-constructed wattle and lime houses neatly whitewashed, having doors and glazed windows. (Methodist Church of Australasia, Letters Received, March 1, 1873)
Reverend Fletcher recognized and continued to use the power of example and intergroup competition to effect changes in construction practices when he returned to the island after Brother Osborne's departure. Following a severe hurricane that devastated crops and destroyed buildings in 1874, he wrote from Rotuma:
The people see the need of better houses, and will gradually I think use stone instead of the plaited cocoa-nut leaves, or even lime. I have just completed a stone room for myself, which will be invaluable as a refuge, should my family need one in another storm, and meanwhile I have a capital study. It serves too for the weekly meetings of my teachers. The building is about twenty one feet by fourteen feet inside. The walls are seven feet high from the floor and twenty inches thick. It is my first attempt as a mason-& may it be my last! It is the first building entirely of stone on the island. I was induced to undertake it partly to encourage the people to let the roofs of their chapels rest directly on the walls. Doubting the security of this arrangement, they preferred to erect the whole framework of the building, and then fill in between the posts with stone and mortar. I have prevailed on the Noatau people, amongst whom I reside, to leave the posts they had prepared, and they are now putting up a new chapel of stone fifty feet by thirty. The height will be about eleven feet. . . . And as the power of rivalry is strong amongst the chiefs, the erection of one good stone place of worship may result in the erection of many more. (Methodist Church of Australasia, Diary of Reverend Fletcher, October 27, 1874)
Brother Osborne was caustic in his assessment of the effects of the work of the two French Catholic priests on the island, in part because the Catholics did not put the same emphasis on changing the domestic conditions of the Rotumans:
It is painful to be compelled to state that Roman Catholicism in Rotumah is really no better than heathenism. It does not raise the people socially or morally; their houses and their persons are nearly as filthy as ever they were. (Methodist Church of Australasia, Letters Received, March 1, 1873)
The Catholic priests, rather, focused on building two huge churches and school complexes on the island. Virtually all the materials for the churches-including wood, stained glass windows, altar, statuary, bells, and even gargoyles for the clock tower-were imported from France. The building process took decades. The priests involved the local people in the construction and decoration of these buildings, thereby teaching them new skills. By 1938 the British resident commissioner, A. E. Cornish, reported:
The new school and dormitory at the Rotuman Catholic Mission, Sumi Station, is now nearing completion and a very worthy and solid building it will be. When completed this will be the best building in Rotuma, even the churches, as buildings, cannot be compared with it. The Sumi Mission school offers more opportunities to boys than ordinary school lessons. Most of the boys turned out by this school are good carpenters and have a good knowledge of cement work, engines etc. The girls at these mission schools do excellent needle-work, frequently gaining prizes at the Suva Show. (Rotuma District Office, Outward Letters: Annual Report of 1938, 9)
In the following year's report Cornish noted the completion of electrical wiring, painting, and building of cupboards at the school, commenting with pride, "Any visitor would be amazed to find such a building in an isolated island such as this" (Rotuma District Office, Outward Letters: Annual Report of 1939, 7).
Environmental, Social and Economic Impetus to Change
The work of the missions-both through inculcating their values about appropriate housing and through teaching construction skills-undoubtedly influenced Rotuman aspirations for European-style housing. But the preference Rotumans demonstrated for new housing styles may have its roots less in ideology than in practical response to opportunity. A number of factors combined in the late 1800s to make a switch to new house styles both possible and desirable.
Rotuma is periodically subject to hurricanes, often necessitating the reconstruction of buildings islandwide. Unfortunately, the supply of thatch is frequently depleted at the very time it is most in demand. A hurricane in 1874 destroyed virtually all the houses on the island (Boddam-Whetham 1876, 262), and replacement thatch was scarce (Rotuma District Office, Outward Letters: November 24, 1884). The resulting housing crisis may have been aggravated by the behavior of the victors in one of a series of religious wars about the same time. The Methodists reportedly burned houses belonging to Catholic and "heathen" Rotumans (Forbes 1875, 242), although such behavior is specifically denied in the accounts given to Gardiner (1898, 470) some twenty years later.
When Europeans introduced the use of lime (soroi), made from burnt coral, as a building material in the 1870s, Rotumans greeted the innovation enthusiastically. At first they plastered it over their thatched walls, then began to build new stone houses (ri hafu), plastered inside and out with lime (Gardiner 1898, 435). By 1884, Resident Commissioner W. M. Gordon reported that "stone-and-lime houses which are well built and accurate in dimensions, are rapidly taking the place of the present thatch houses" (Rotuma District Office, Outward Letters: November 24, 1884).
To the extent that the new style houses were built of local resources, obtaining and preparing materials and erecting structures remained processes dependent on kainaga assistance. Like Rotuman thatched houses, limestone houses were built on named foundations and embodied in tangible form the caring support of the relatives who contributed to their existence. However, they also set a new standard for what constituted a good house. According to Allen (1895), the Rotumans building stone houses used "wooden doors, and windows of European manufacture." They began to incorporate other imported materials as well, such as cloth curtains and corrugated iron roofs. Acquired through barter or purchase, these materials reflected the increasing participation of Rotumans in the market economy through sailing and copra trade as well as widening exposure to alternatives for house styles and furnishings. 
The proportion of ri hafu on the island gradually increased over the next several decades. In a 1948 report commissioned by the colonial government, J. W. Sykes wrote that "most of the houses are built of stone cemented with a mixture of coral lime and sand and covered with a roof of sago palm leaf thatch" (Sykes 1948). He noted that there were also many European-style houses with wooden walls and iron roofs, although these were not well maintained. One would gather from his report that there were few, if any, houses with thatched walls on the island. But just a few years later, District Officer H. S. Evans (1951, note 25) provided a numerical assessment of housing types that indicates thatched structures had persisted to some extent. He reported, "Rather over one third of the houses are attractive cottages of coral lime concrete, brilliantly white with lime wash; rather less than one third are Rotuman houses with sago leaf walls; twenty-four percent are timber houses and the remaining nine per cent of corrugated iron." Most roofs were still thatched, with no more than twelve percent of the houses having iron roofs.
By 1966, according to a report prepared by the Rotuma Council of district chiefs and representatives (see table 2.2), more than half the houses had stone or cement walls (both called ri hafu). Thatched houses (ri ota) had decreased to less than one-fifth, houses with iron walls (ri pota) had increased to 18 percent, and only 13 percent were timber houses (ri 'ai). Fifteen years later, in 1981, the Rotuma Council reported that 83 percent of houses had stone or cement walls. Wooden and iron-walled houses constituted 10 percent and 8 percent of island houses, respectively. These changes, and especially the fact that there were virtually no Rotuman-style thatch houses standing, were due in large part to the 1972 hurricane named Bebe and the relief program that followed.
Rotuma House Styles, 1951-1989
Hurricane Bebe destroyed or damaged most buildings on Rotuma. Afterward, under the provisions of a government disaster relief program, Rotumans were given small loans (averaging about F$274) in the form of materials, typically including six bags of cement for a house foundation, eight galvanized iron pipes for supports, timber for rafters, roofing iron, and nails. The New Zealand Army came to Rotuma to assist with the rebuilding effort, and brought the materials. The rafters were cut and assembled at one site, then loaded onto a truck with the other materials and delivered to sites around the island. A model house was built in one district, with two or three men from each district assisting. These men then worked along with one soldier, assigned as foreman, to direct construction by eight-person teams in their own district. After pouring the foundations, the teams placed iron posts upright in the cement to act as roof supports, then erected the rafters. People were left to choose and build their own walls out of whatever material they could afford and obtain. 
The construction teams competed to see how fast they could build the basic structures. The work of the New Zealand Army and their Rotuman assistants has now assumed legendary status on the island: During a period of twenty-one days, I was told, they built 302 new housing units. As can be seen from the house counts in table 2.2, this represented a significant proportion of the dwellings on the island.
Besides the obvious physical differences, this massive reconstruction effort provided opportunities for other kinds of change. Some families chose not to rebuild their houses on kin group house foundations.  According to the 1989 survey, only 58 percent of island households were located on fuag ri. Although the majority of Rotuman homes are still built on kainaga land,  those located away from fuag ri may be less subject to claims by other kainaga members. Those investing time and materials in a more permanent structure may have been hoping to ensure its being passed on to their own offspring.
Hurricane Bebe and subsequent government aid provided significant impetus to housing changes on Rotuma. However, the overall trend toward more elaborate, individually owned housing is sustained to a great extent by the outmigration of Rotumans to paid positions abroad, and the cash and imported materials these migrants send back to the island.
As mentioned above, there has been a dramatic increase in the proportion of Rotumans living in Fiji over recent decades. Although the population remaining on the island remained fairly stable from 1956-1986, average household size decreased from 7.4 to 5.8 persons. Much of this can be attributed to a marked increase in the number of households with one to three persons. While in 1960 Howard found that such small households made up only 11 percent of Rotuman households, in 1989 almost 30 percent of households fell into this category. At the same time, the percentage of households on the island composed of ten or more people dropped from 17 percent in 1960 to only 8 percent in 1989 (figure 2.3).
Cars & trucks
Data obtained from 1989 survey of 414 houses conducted by Jan Rensel and Alan Howard
With more and more Rotuman houses equipped with Western-style furniture, mat making assumes a smaller portion of women's responsibilities. Although mats remain highly important for ceremonial exchange, and commonly are used as floor coverings or beds, women reportedly spend less time plaiting mats than they used to, and young women often do not learn how to plait mats at all. When I asked what women are doing instead, people suggested they were spending more time looking after their houses. Respondents to a 1989 survey reported activities such as sewing curtains and bed sheets, making doormats, and crocheting doilies, not to mention washing and ironing household linens. Much attention went into the appearance of houses, with borders of colorful bougainvillea and croton bushes planted outside and containers of fresh or plastic flowers, pictures and other ornaments decorating indoor spaces. The care of houses seems to have assumed greater importance in Rotuman perceptions than in 1960 (Howard, personal communication). Although some activities, such as needlework, are pursued in social settings such as women's groups, more of the house-related work done by women today is done individually.
Valuation and Support of Relationships
Changes in housing on Rotuma obviously go far beyond physical structures. The decision to invest in a new house has significant implications for a household's relationships in the Rotuman community. In choosing to build or extend with imported materials, one frequently is emphasizing ties with migrants over those with local kainaga. Participation in the work process is limited, including fewer women and only men with particular skills. While a house built of local materials by a large cooperating group stands as a constant reminder of their care and support, one constructed by few, paid workers embodies correspondingly less social meaning.
Subsequent activities are also affected. With the increasing practice of giving money for help and materials, reciprocal assistance is downplayed. With changes in the form and furnishing of dwellings, opportunities for shared activities between nearby households are diminished. For households with fewer members, correspondingly greater attention may be devoted to supporting relationships with off-island relatives by sending produce and hosting visits.
Location and Permanence
Since Rotumans reckon rights to house sites rather than to specific structures, a house built on family land not on a traditional foundation may be reserved for one's children without contention. The construction of permanent houses on fuag ri will necessitate some renegotiation of criteria for claims to the site. It appears that capital investment in a house is being recognized as sufficient justification for a lineal family group to remain on a fuag ri. This has the added implication of strengthening the claims of immediate descendants, by their continuing presence, to not only the site but associated garden lands and title, if any.
The shift to permanent buildings itself has implications for Rotuman dispute management. Avoidance is one of the major strategies for dealing with conflict. A serious falling-out may result in one party's relocating to avoid contact with the other. One such instance arose in 1989, when a household tore down their thatched dwelling and rebuilt it at another location because of a disagreement over claims to the first site. This option is practicable for people with iron or wood houses, but out of the question for those with cement buildings. People may be able to get away temporarily by visiting relatives elsewhere on the island, in Fiji or abroad, but eventually they must return or face having to give up a sizable investment of cash, labor, and materials. Even more difficult are disputes over land claims in which someone else attempts to force a household to leave. Bad feelings are exacerbated by the specter of losing not only the land but a permanent house and the work and relationships it represents.
Wealth, Rank, and Social Merit
Rotuman concepts of what constitutes a good house have been shaped by missionary teaching and example, experiences with other outsiders, and concerns for practicality and convenience. Whatever its genesis, the predominance of European-style housing bears witness to a valuing of imported over indigenous models. The switch to concrete structures after Hurricane Bebe was impelled not only by practical concerns and a desire for stronger materials, but also by a pursuit of European goods for status purposes. By embracing Western-style goods as status markers, Rotumans perhaps inadvertently have contributed to changes in social relationships.
Over the past few decades, according to some observers, a Rotuman's house has become the "measuring-stick whereby one gauges people's wealth and status" (Plant 1991, 205). Prior to cession in 1881 there was little material difference among Rotuman houses in style and furnishings. Chiefs' houses were distinguished primarily by their larger size, which reflected chiefly responsibility for hosting visitors. In the past chiefs could call on community labor to build their houses, but today if they want something other than a thatched dwelling they are in the same position as everyone else: Materials, and to some extent labor, cost money. A related change is an apparent decline in the custom of claiming the particular house site that goes with a title when a person is made a chief. When a man appointed to a certain subchiefly title moved back to Rotuma from Fiji in 1988, the household head staying in a cement house on the site affiliated with that title adamantly refused to allow the returnee to move there. The new subchief subsequently built a thatched house on other kainaga land.
At the same time, for some, Western-style houses may be becoming increasingly important for establishing claims to chiefly titles. In 1988, during a discussion of qualities to look for in a candidate for chieftainship, one Rotuman suggested to me the following criteria, in this order 91;emphasis added]:
The inclusion of housing in such formulations was not noted during previous anthropological research on the island in 1960 (see Howard 1970). A negative example is the case of one district chief who in 1990 was subject to criticism for having a humble thatched dwelling rather than a "proper house" in which to entertain visitors. At present, the houses of most district chiefs are cement structures, but neither the largest nor the most imposing in their districts.  Rather, people with higher earned incomes, or financial support from off island, command the resources to develop elaborate housing.
A Good House--and A Good Provider
Evaluation of social merit aside from rank considerations seems also to have been affected by increased access to Western-style housing. Customarily, a Rotuman's ability to provide an abundance of food, primarily garden produce, has been of central importance in evaluating social merit. Recently, however, there is some suggestion this measure has been eclipsed by one's ability to provide a Western-style house. Wilson Inia, Rotuma's first senator to the Fiji legislature, said in a 1974 speech in support of the savings and house loans programs of the Fiji National Provident Fund:
One of the great responsibilities of a father to a family is to provide a house while he is alive, or if he has passed away, to leave behind sufficient funds for the mother or the children to build a house. That is good advice to a Rotuman whether he be in Suva or in Rotuma. Any father who cannot provide that is a bad father. (Parliamentary Debates, October 14, 1974)
I was told in 1988 that there is a Rotuman saying: "Nono ka ri lelei, ma 'inea ne hua' lelei." 'When the house is good, you know the occupants [those who look after it] are good.' But this may not represent as much a disjunction from former bases for attributing merit as it may first appear. The term hua'i (shortened to hua' in this context) connotes the work of caretaking. Rotumans who are long-term residents on the island, and are aware of the social histories of buildings, are more likely to distinguish between merely having a nice house and having done the work to procure the materials and build one. Not taken in by appearances, those who know whose work is represented can judge houses much as they evaluate food production and contributions.
Social pressure does not deter Rotumans from trying their best to build and furnish their preferred house in whatever ways are open to them. But not all Rotumans make the same choices. Over the past decade, the advent of two-story houses on the island represents one extreme. The man who built the first such house was subject to criticism from others for his ostentatious display, but others have since begun to follow his example: The 1989 survey turned up six houses with two stories. These have been built with migrant or returnee money and represent a valuing of comfort and status (defined in urban wage-earner terms) over fitting into the community. In fact, for people with less urban experience, such elaborate houses have a distancing and intimidating effect. At the other end of the spectrum are the examples, mentioned above, of the two migrant Rotumans who built thatched houses when they returned to the island to accept chiefly titles. Their decision to do so may reflect other considerations, such as a desire to limit monetary investment in imported materials until they could see how the new positions would work out. But by electing to build traditional Rotuman houses with the help of the people they came to lead (and serve), these new chiefs also gave priority to reconnecting with the community through the familiar and time-honored practice of reciprocal labor.
This paper is based on a review of historical documents, including the field notes of previous ethnographers, and on research I conducted with Alan Howard on Rotuma during six field trips, ranging from a few weeks to six months, between 1987 and 1994. My research was sponsored in part by a Fulbright Grant administered by the Institute for International Education. Oral versions of this paper were presented at the University of Hawai'i and the University of the South Pacific. I am grateful to those who responded with comments and suggestions, especially Dr. Vilsoni Hereniko, Paul Vaurasi, Mrs. Elisapeti Inia, and Prof. Asesela Ravuvu. This chapter is a reorganized and expanded version of an article entitled "Housing and Social Relationships on Rotuma" that appeared in Rotuma: Hanua Pumue (Precious Land) (Fatiaki et al. 1991).
 For a guide to Rotuman pronunciation, see Churchward 1940, 13. [back to text]
 By the end of 1995 both cooperatives had ceased doing business. While this sizably reduced the number of wage positions on the island, opportunities for entrepreneurs have increased, and a number of individual trading enterprises have sprung up. [back to text]
 One Fiji dollar is worth approximately sixty-seven cents in U.S. currency. [back to text]
 Elisapeti Inia instigated the development of curriculum materials for teaching Rotuman language, oral traditions, and custom in schools in Rotuma and Fiji. Her materials were recently adopted and expanded under the auspices of the Curriculum Development Unit of the Fiji Department of Education, with Mrs. Inia's guidance and involvement. [back to text]
 Through their own experiences abroad as sailors, and the example of visiting Europeans, Rotumans were exposed to and adopted a number of innovations in furnishing their houses. For instance, a Mr. Emery, former mate of an English whaleship who left that position for health reasons, settled in Rotuma around 1829 and built a wooden house on the offshore islet of Uea. He had English furniture, cooking utensils, and pictures on the walls. Emery married a Rotuman woman, and lived on Uea with about sixty other Rotumans who reportedly treated him as their chief (Cheever 1834-1835). Another sailor from a whaling ship, visiting in the early 1850s, noted that brightly colored curtains were used to screen the sleeping areas of a large house he and his mates visited. He surmised these had been traded by some whaling captain for hogs and other provisions (Haley 1948, 258). [back to text]
 Due to lack of funds, and competing demands for what money some families did have available, many houses took years to complete. [back to text]
 Paul Vaurasi, a Rotuman who worked for many years in the Fiji government's Department of Public Works, noted that new fuag ri are sometimes constructed in order to create level ground on which to build. [back to text]
 Of the 414 households responding to the 1989 survey, 306 (74 percent) said their houses were located on kainaga land. Other possibilities were hanua togi, land owned outright as a result of purchase from other Rotumans (9 percent); hanua na, land owned outright as a gift (2 percent); hanua pau, land owned outright by those residing there, the only claimants as a result of attrition (3 percent); hanua haisasigi, land belonging to siblings (3 percent); lands belonging to government (2 percent) and church (3 percent); and no information given (4 percent). [back to text]
 As in English usage, there are several euphemisms for the toilet in Rotuman, including ri mea'me'a 'little house,' ri la'oaga 'house for going,' and fa'u 'out back' as in ia la' se fa' 'he went out back.' When Howard conducted his fieldwork on the island in 1960, outhouses were located either inland (the back, according to Rotuman orientation) or at the end of a wooden pier leading from the beach to beyond the high tide mark (Howard 1970, 31). [back to text]
 When roofing iron is replaced on dwellings, the old iron is often reused for kohea or other outbuildings. [back to text]
 I am grateful to Paul Vaurasi for pointing this out. [back to text]
 Howard reports that in 1960 only in a few villages did the boys build their own sleeping house; more often they used a structure that was temporarily available, or went to the home of an older single or widowed man (Howard 1970, 66). During my fieldwork in the late 1980s, I observed that while some slept in the home of their parents, many of the young men in Oinafa village took their mats and mosquito nets to the community hall and slept there instead. [back to text]
 Responses to the survey may reflect rhetoric more than reality, especially in cases in which people did not know the details of their dwelling's history. [back to text]
 Some Rotumans pay skilled laborers cash for their work, recognizing that government and other organizations pay them for doing this type of work. In other cases, such as a flurry of house renovations in Oinafa prompted by a large celebration in 1989, skilled assistance is in great demand and short supply; people essentially competed for the workers' time by offering F$10 to F$12/day. [back to text]
 Interestingly, I observed that when a Rotuman household hosted a large group inside the dwelling, for instance for a small ceremony or a prayer meeting, they often pushed aside chairs and sofas or removed them from the area, spreading mats on which people sat. Although household members used their furniture on an everyday basis, with the arrival of even casual visitors everyone frequently ended up sitting on the floor. [back to text]
 An exception is the case of the chief in the highest ranking district on the island. A large guest house with a high roof and commanding aspect was constructed to house visiting VIPs during the 1981 celebration of the centennial of the cession of Rotuma to Great Britain. After a new district chief was installed in 1983, he claimed the guest house as his residence. [back to text]
Allardyce, W.L. (1885-6) Rotooma and the Rotoomans. Proceedings of the Queensland Branch of the Geographical Society of Australasia. 1st sets: 130-144.
Allen, William (1895) Rotuma. Report of Australasian Association for Advancement of Science. Pp. 556-579.
Bennett, George (1831) A recent visit to several of the Polynesian Islands. United Service Journal 33:198-202, 473-482.
Bertram, G. and R.F. Watters (1985) The MIRAB economy in South Pacific Microstates. Pacific Viewpoint 26:497-519.
------ (1986) The MIRAB process: Earlier analyses in context. Pacific Viewpoint 27:47-59.
Boddam-Whetham, J.W. (1876) Pearls of the Pacific. London: Hurst and Blackett.
Cheever, George N. (1834-5) Log of the ship Emerald. Pacific Manuscripts Bureau frame 31.
Churchward, C. Maxwell (1940) Rotuman Grammar and Dictionary. Sydney: Australasian Medical Publishing Co.
Eagleston, John Henry (1832) Log of the ship Emerald, Vol. 3. Archived in the Peabody Museum, Salem, Massachusetts.
Evans, H.S. (1951) Notes on Rotuma. Typescript.
Fatiaki, Anselmo et al (1991) Rotuma: Hanua Pumue (Precious Land). Suva: Institute for Pacific Studies, University of the South Pacific.
Forbes, Litton (1875) Two Years in Fiji. London: Longmans, Green, and Co.
Gardiner, J. Stanley (1898) Natives of Rotuma. Journal of the Royal Anthropological Institute 27:396-435, 457-524.
Haley, Nelson Cole (1948) Whale Hunt, the Narrative of a Voyage 1849-1853. New York: Ives Washburn, Inc.
Hocart, A. M.1913 Field notes. Archived in Turnbull Library, Wellington, New Zealand.
Hooper, Antony and Judith Huntsman (1973) A Demographic History of the Tokelau Islands. Journal of the Polynesian Society 82:366-411.
Howard, Alan (1970) Learning to Be Rotuman. New York: Columbia Teachers College Press.1960 Unpublished field notes.
Lesson, René (1838-9) Voyage Autour du Monde sur"La Coquille". Paris: Pourrat Fréres.
Lucatt, Edward (1851) Rovings in the Pacific, 1837-49 by a Merchant Long Resident in Tahiti. London: Longman, Brown, Green, and Longman.
Methodist Church of Australasia, Department of Overseas Missions (1855-1879) Letters Received, Fiji-Rewa, Rotuma, Ovalau, Bau & miscellaneous. Unpublished documents, file 98. Archived in Mitchell Library, Sydney.
------ (1873-1876) Diaries and other personal papers of Reverend William Fletcher, Noatau. Unpublished documents, file 329. Archived in Mitchell Library, Sydney.
O'Meara, Timothy (1986) Why is Village Agriculture Stagnating? A Test of Social and Economic Explanations in Western Samoa. Ph.D. Dissertation, University of California-Santa Barbara.
Osborn, Joseph W. (1834-5) Log of the Ship Emerald of Salem, Captain John H. Eagleston. Pacific Manuscripts Bureau microfilm 223, frame 151.
Parliamentary Debates: Senate (1974-1984) Fiji Legislative Papers. Suva: Government Printing Office.
Plant, Chris (1991) The development dilemma. In Rotuma: Hanua Pumue (Precious Land) Fatiaki, Anselmo et al. Pp. 204-226. Suva: Institute for Pacific Studies, University of the South Pacific.
Rensel, Jan (1993) The Fiji connection: Migrant involvement in the economy of Rotuma.Pacific Viewpoint 34 (2): 215-240.
Rotuma District Office (n.d.) Outward Letters. Archived in the Central Archives, Suva, Fiji.
Severance, Craig (1976) Land, Food and Fish: Strategy and Transaction on a Micronesian Atoll. Doctoral dissertation, Department of Anthropology, University of Oregon.
Shankman, Paul (1976) Migration and Underdevelopment: The Case of Western Samoa. Boulder, Colorado: Westview Press.
------ (1992) The Samoan exodus. In Contemporary Pacific Societies: Studies in Development and Change, edited by Victoria S. Lockwood, Thomas G. Harding and Ben J. Wallace. Pp. 156-170. Englewood Cliffs, NJ: Prentice Hall.
Sykes, J.W. (1948) Confidential Report on Rotuma. The Secretariat, Suva, Fiji.
Taito, F. (1940). My Own Story. Sydney: Methodist Overseas Missions of Australasia, Literature Department.