- 199

Over the past decade there has been renewed interest in the history of population change in the smaller island territories of the south Pacific. The growing body of evidence and continuing debate clearly demonstrate that the demographic responses of island populations to protracted contact with the Western world were far more varied than was formerly believed. It is not disputed that there was some overall decline in the population of the south Pacific region during the 19th and early 20th centuries. However, when individual countries, islands and communities are examined, their courses of change appear to differ markedly. McArthur's (1967) review of the populations of five countries in Polynesia and Melanesia demonstrates this clearly at the national level; a more recent collection highlights variability at community and island levels (Carroll 1975). In this article we demonstrate similar variation at both local and regional levels in two island groups, Kiribati and Tuvalu, which have received little direct attention in the literature (Fig. 1). Aside from a brief historical introduction to a census in 1947 (Pusinelli 1947:1-3), the only published article dealing specifically with demographic change in this area is Newton's (1967) short critique of some early population estimates for Tuvalu.

An interpretation of demographic developments between 1850 and 1900 forms the substance of this article. These 50 years cover the period between the first reasonably complete run of estimated island populations for Kiribati and the first official census in 1895. Considerable population decline is sug-

- 200

FIGURE 1

Illustration

- 201

gested when mid-19th century estimates are compared with numbers subsequently recorded in census enumerations in the late 19th and early 20th centuries. In our view, the magnitude of depopulation seems greatly exaggerated by this simple comparison. Our primary objective is to re-evaluate the early estimates, some of which have become “enshrined in the literature and thereby endowed with an undeserved reality”, to echo Norma McArthur's (1970:1101) words for pre-census estimates in other parts of the Pacific.

To provide some context within which to place the numerous estimates of island populations cited in the paper, we begin with an evaluation of our data base. This is followed by a short description of environment and society in Kiribati and Tuvalu, in which particular emphasis is placed on customs and events known to have influenced the course of demographic change before protracted foreign intrusion. Subsequent trends in fertility, mortality and migration between 1850 and 1900 are examined in the third section where we draw on a wealth of unpublished information contained in mission records, colonial administration archives, and the diaries of traders, sailors and adventurers of the period. The numerous population estimates mentioned in these sources are discussed in the last section. Details of the estimates, together with our assessments of possible mid-century populations for different parts of the two island groups are contained in Tables 5 to 13 (following article text). We wish also to draw attention to an interesting comment (Note 26) on our assessments of likely populations for Kiribati which was received from H. E. Maude after this manuscript was submitted for publication.

In 1892, when Kiribati and Tuvalu became British Protectorates, the indigenous populations were estimated to total about 29,500 — 26,400 Micronesians and 3,100 Polynesians respectively. Yet 30 years earlier Richard Randell, a trader resident on Butaritari, claimed that the Micronesian inhabitants of Kiribati numbered between 50,000 and 54,000 (Table 1). Even higher estimates were recorded in journals of the United States Exploring Expedition in the early 1840s where figures of 60,000 and 85,000 Micronesians are cited (Wilkes 1845:105; Hale 1846:93). For the Polynesians in Tuvalu, a pre-contact population of as many as 20,000 has been suggested by Louis Becke, a visitor to this part of the Pacific in the 1880s (Pusinelli 1947:1). These estimates by Randell and Becke have been quoted repeatedly in official documents and publications as probable populations for Kiribati and Tuvalu before extensive contact with Europeans. A dramatic decline in numbers is suggested when such figures are compared with the populations Davis considered to be present on each island in 1892. As Newton (1967) has already demonstrated for Tuvalu, and as we show below for Kiribati, this view of severe population decline cannot be substantiated.

- 202

Regular contact with Europeans in this part of the Pacific was initiated by whalers seeking the favours of local women and occasionally trading for coconut oil and supplies in the 1820s and 1830s (Maude and Leeson 1968:235). Itinerant traders followed and were in turn succeeded by a few beachcombers and resident traders. By the early 1860s there were about 50 Europeans resident in Kiribati (Maude and Leeson 1968:239). It was during this decade that the London Missionary Society (LMS) began operations in Tuvalu. In 1870 they extended their activities to southern Kiribati. In missionaries' touring reports, as well as in the diaries of early traders and visitors, population estimates for specific islands are frequently mentioned. Over 350 counts and guesses of numbers on different islands in the two groups during the second half of the 19th century can be found in such sources.

It was not until 1892, however, that figures for all islands in Kiribati and Tuvalu were given for a particular year in the same source (Tables 1 and 2). Between 1852 and 1871 there were at least three estimates by island in Kiribati, but for the subsequent 20 years there are no complete runs for the 16 inhabited islands. As far as Tuvalu is concerned, population summaries by island for specific years before 1892 are more frequent, especially in records of the London Missionary Society. A selection of the more complete estimates for the two areas is shown in Tables 1 and 2. Figures obtained in the first official census in 1895 are also given, together with those cited by Captain Tupper, of HMS Pylades, in 1901. Although there are only five complete runs for Kiribati between 1850 and 1901 it appears that, for the group as a whole, the indigenous population was declining during the latter half of the 19th century. In the case of Tuvalu, however, missionary estimates between 1876 and 1887 plus official counts in the 1890s and 1901 point to steady increases in numbers.

Evaluating the Evidence: An Approach

Although there seems little doubt that Randell's populations for many of the islands in Kiribati are exaggerated, we undertook the task of reassessing early estimates with a number of reservations. One approach would be just to present the figures in tabular form and leave it to the reader to make an appraisal without suggesting any alternatives to Randell's estimates. Undoubtedly this would be the safest strategy to adopt, bearing in mind McArthur's (1970: 1101) caution that “critical evaluation of the early historical estimates of the size of a nonliterate population newly contacted by literate people discourages the acceptance of much of the information recorded, and even that which is acceptable raises more questions than can now be answered”. However, we consider some evaluation must be attempted and, where necessary, some alternatives offered, because there is still a tendency for Randell's figures to be cited quite uncritically by administrators and academics alike. Although we

- 203

- 204

- 205

can offer nothing more than another set of “informed guesses”, we hope the attempt will lead to more critical appraisal of 19th century population estimates and demographic trends in this part of the Pacific.

It is not our intention to discuss all the population figures we have for different parts of the two groups. In Tables 5, 7, 9 and 12 virtually all of those at our disposal are listed, along with their sources. Some represent nothing more than an itinerant trader's or visiting naval commander's assessment of possible numbers on the basis of size of island and the frequency with which villages were observed, but the majority are counts and estimates made by resident missionaries or traders. As far as the estimates for 1841 are concerned (cited in Wilkes and Hale), it should be noted that these are very unreliable. In some cases (e.g., for Abemama, Kuria and Aranuka) the figures are mere extrapolations of a resident beachcomber's guess at the number of warriors he observed on a particular occasion; in others they are estimates of atoll populations on the basis of number of residents assumed to be present in one village which was visited, multiplied by a rough count of other settlements made as the vessel sailed through the lagoon. The prevalent settlement pattern was a continuous scatter of houses or small hamlets along a lagoon foreshore, and it would not be difficult to gain an impression of much higher population densities than were really there. When numerous clusters of houses were seen this did not necessarily mean that an island was densely populated, as there was generally little depth to the ribbon settlement pattern.

Another reason for the possible overestimation of the total size of an atoll's population by a visitor was a tendency for residents to flock to any point where a foreign ship dropped anchor. Because navigation of many lagoons was very difficult, visitors were far less mobile than the inhabitants, who could regroup at each of the vessel's landing points. Finally, it should be noted that many figures cited in mission reports could relate to de jure rather than de facto populations. The usual response to a question addressed to a chief or headman about the number of people in his village probably would have been an estimate of all those who “belong” to that place whether they were resident, temporarily absent, or had been away for some years. This problem of the relevant population base becomes critical from the 1870s when there were absentees from all islands overseas; unfortunately, rarely is any indication given about the status of the population to which reference is made in most mission reports. This factor, coupled with extensive inter-island mobility (discussed below), could account for some fluctuations in numbers which exist in the record of estimates for islands in southern Kiribati and Tuvalu.

On the basis of available data it is possible to suggest quite feasible populations for Tuvalu before labour recruiting began in 1863. We would acknowledge, however, that there is a much greater element of doubt about - 206 our figures for Kiribati and for this reason a range for the population of each island is suggested, rather than a single number. Also provided are rates of net population gain or loss between 1860 and 1900, along with the crude population densities based on total land areas. The latter are for illustrative purposes only; we have not attempted to derive “maximum” densities or theoretical “carrying capacities” for the different islands although mention should be made of recent research in this area on atolls (Bayliss-Smith 1974, 1975). Bayliss-Smith's (1974:287) average crude density for five atolls to the north of the Solomon Islands before prolonged European contact (295 persons per square kilometre which was derived from a range between 230 and 680 per square kilometre) is likely to be higher than that which many of the drought-prone southern and central islands in Kiribati could support. This is a subjective judgement; our data do not lend themselves to more precise carrying capacity estimations.

Compared with larger volcanic islands to the north and south, atoll and reef island ecosystems in the central Pacific have quite limited potentials for supporting human populations. All the islands of Kiribati and Tuvalu are of coral formation with elevations of less than five metres. None of those in Tuvalu exceeds six square kilometres. While the islands in Kiribati are generally larger, only three have land areas in excess of 20 square kilometres and all are less than 40 square kilometres. In addition to small total size, the land area on atolls is usually in the form of a discontinuous string of islets along an extensive reef. As a result of morphology and geology, fresh ground water supplies are restricted to shallow sub-surface lenses and the calcareous soils are at the margins of fertility. Except for islands in the northern and southern extremities, rainfall is variable and often unreliable. Drought is a common hazard in southern Kiribati especially, and has been reported as quite a significant cause of mortality at times. 17 Other environmental hazards such as hurricanes and abnormally heavy seas also affect islands in Tuvalu occasionally, causing extensive damage to coconut groves and dwelling places. 18 The range of plant species which can survive on these small islands with their skeletal soils, irregular rainfall and limited ground water is severely restricted. The coconut palm and pandanus tree (Pandanus pulposus) are ubiquitous; other basic food crops — taro and Cyrtosperma chamissonis, a coarse tuber, require careful cultivation in specially prepared pits. Breadfruit, which may not have reached some islands in southern Kiribati until the latter part of the 19th century, bananas and pawpaw have been added to the local diet.

Population Control before Colonial Intrusion

It is clear from oral traditions and references to contemporary customs by the early missionaries that the Micronesian inhabitants of Kiribati and Poly- - 207 nesians in Tuvalu pursued, either intentionally or otherwise, a policy of population control well before European contact. Abortion is reputed to have been common practice and a variety of methods have been reported. 19 Infanticide was also practised, especially in Tuvalu. 20 The extent of abortion and infanticide, and their effects on family sizes is difficult to determine from available evidence, but it would appear that these methods definitely contributed towards controlling growth in small island populations.

There were other constraints on fertility levels as well — especially various customs which favoured late marriage by men and lengthy gaps between pregnancies for women. Arthur Grimble (1921), administrator and amateur anthropologist, described in some detail gruelling manhood initiation rites which tended to delay the age of marriage for males in Kiribati until their late 20s. Men could seek sexual gratification before marriage from among the nikiranroro, women who were known to have had pre-marital sexual experiences. Virginity among brides was generally demanded, however, and there was considerable control over the social behaviour of young women. Grimble mentioned that females tended to marry much younger than men, usually at 14 to 15 years of age. Polygyny is reported to have been practised, but on all islands to the north of Abemama only chiefs could have more than one wife (Grimble 1921:27). In Tuvalu virginity was less highly prized than in Kiribati according to Donald Kennedy (1931), another officer in the colonial administration who spent many years studying the customs and languages of the Polynesian inhabitants. Kennedy (1931:308) reported that polygyny was practised in Tuvalu, but adds the rider that “it was as much as the ordinary man could do to provide for one wife and her offspring”.

Both Grimble and Kennedy mentioned that women breast-fed their babies for at least 18 months, during which time intercourse between husband and wife was strongly discouraged. If the mother became pregnant during this period the baby would be weaned immediately; the milk was considered unsuitable for consumption after conception. However, a post-partum sex taboo imposed no great hardship on husbands; in both groups various arrangements ensured that men could find sexual gratification readily without offending cultural norms. In Kiribati the custom of eiriki ‘wife's sister exchange’ gave a husband access to his wife's sisters who would play the part of wife until the baby was weaned (Grimble 1921:36). In Tuvalu the custom of moetotolo (Kennedy 1931:304) permitted men to seek sexual favours among single as well as married women. 21 Casual adulterous relationships were not regarded as punishable offences for either sex in Tuvalu. Although some of these reported practices do not suggest very obvious control over fertility levels, their combined effect would constrain birth rates.

These islands in the central Pacific were relatively disease-free, and high fertility to compensate for considerable infant mortality does not seem to have been an established cultural response. The only widespread endemic dis- - 208 eases were yaws and, in Tuvalu, filariasis. While both could be very debilitating, they were seldom fatal and premature death of an adult through sickness was an unusual occurrence. 22 Populations were, in general, more healthy before European contact than after; the first explorers frequently commented on the good health of the inhabitants (Pine 1972:187).

An important cause of mortality among adult males in Kiribati was intergroup warfare. With the exceptions of Makin and Butaritari, fighting between different chiefly clans during the 19th century in islands north of and including Abemama was virtually endemic and probably served as a periodic check on population growth. Rivalry between groups on some islands, such as Tarawa and Abaiang, can be traced back for generations and disputes in this area continued until the Protectorate was established in 1892. In Tuvalu there are few traditions of inter-group fighting during the 19th century, although one feud between Funafuti and Nukulaelae about the 1850s is referred to by Hedley (1896). 23 Another dimension to mortality associated with war was cannibalism, which Grimble (1933:60-1) notes was practised occasionally in Kiribati. It was more common, however, to regard the biting of an enemy's eyes as sufficient humiliation for the dead, and there is no record of deliberate raids to obtain human flesh for feasts such as those that were repeatedly mentioned in writings of missionaries in northern and central Fiji about 1850.

Population Movement between Islands

Population redistribution was a significant demographic consequence of many wars in the central islands of Kiribati. During and after a major clash the vanquished might flee or be driven from their lands. Members of the United States Exploring Expedition, in Kiribati in 1841, reported that fugitives who reached a haven might attempt to reconquer their islands or plot to overthrow their hosts; they specifically mention a case where the latter strategy was adopted unsuccessfully by a group from Abaiang who sought refuge on Makin in the 1830s (Wilkes 1845:91-2). In addition to changing patterns of population distribution, warfare often resulted in extensive damage to gardens which brought about severe food shortages, especially on the smaller islands. One of the best-documented wars in central Kiribati occurred during the early 1860s (Roberts 1953; Maude 1970); we discuss some of its demographic consequences in the next section.

Inter-island contact was brought about not only through warfare but also by the long tradition of voyaging between islands within Kiribati and Tuvalu as well as to places in other parts of the Pacific. 24 Kiribati sailors were renowned for their skills as navigators, and contact with other Micronesian groups to the north and east was maintained quite deliberately (Lewis 1972). On a more localised scale, links were strongest between selected groups of islands — Maude (1963:9-10), for example, mentions that the people of Buta- - 209 ritari and Makin had a separate dialect and distinctive customs as a result of several centuries of partial isolation from the rest of Kiribati. Intermarriage between people from different islands was also quite common. These various forms of contact encouraged much temporary circular mobility as well as some permanent migration. Islands where populations were reduced by natural disaster, war, or, later in the 19th century, alien diseases or labour migration, were often repopulated by kin and friends from neighbouring islands. In this regard, Hedley (1896:232) notes that when Funafuti's population was severely reduced by recruiters seeking slave labour for Peruvian mines in 1863 “the place of the expatriated natives was largely taken by immigrants from other islands”.

Not all inter-island voyages terminated at the chosen destination. Early in the 20th century the British administration prohibited inter-island travel by sailing canoe because vessels frequently failed to make the expected landing. Maude (1971:xi), citing a late 19th century source, notes that “even in the brief historic period we have 30 authenticated cases of Gilbertese canoes drifting to islands throughout the Western Pacific, from the Carolines to the New Hebrides.... When one adds the many fugitives from inter-island and civil wars and the surplus population compelled by community pressure to migrate or be killed one can gain some conception of the numbers which must have reached some part of Melanesia or [northern] Micronesia, even allowing for the fact that the majority no doubt perished at sea”. Maude (personal communication to B. K. Macdonald) estimates that in some years as many as 1,000 from Kiribati were lost at sea or stranded on alien shores.

Mortality

Alien diseases (in particular dysentery and influenza) were introduced from the earliest years of interaction with Europeans, but there is no oral or written record of group-wide population decline as a result of epidemics. Undoubtedly diseases caused much sickness and discomfort among residents on certain islands at different times, especially after the 1870s when Micronesians in particular began to return after periods in wage employment in other countries. Yet mission records before 1900 rarely make reference to large numbers of deaths due to epidemics. On the occasions that significant disease mortality is mentioned, it is often due to dysentery outbreaks associated with the return of people to their places of residence after a visit to a neighbouring island. 25 However, there are no stories of catastrophic epidemics affecting large areas of the two groups in the early contact period which are comparable to some in Fiji that were remembered and recounted in song and legend throughout the 19th century (Corney, Stewart and Thomson 1896:34-5).

- 210

Although there is a dearth of written evidence of epidemics and massive disease mortality, it is important to appreciate that even a slight increase in mortality levels can have important structural implications for small populations. If children suffer high levels of mortality, then not only does the population experience an immediate decline in numbers, but its future reproductive potential is also seriously affected. Unfortunately, there is no consistent information on the age-selective nature of alien disease mortality during the 19th century in either Kiribati or Tuvalu. Nor are age-sex compositions of island populations specified for different years. Occasionally in mission reports for Tuvalu during the late 1870s and the 1880s a crude classification of the population into “church members”, “candidates”, “school boys” and “school girls” is provided, but these figures do not take account of the “heathen” and very youthful components, and are a poor data base from which to proceed with analysis of structural characteristics of island populations.

In the absence of more precise evidence on mortality from epidemics it is probably justifiable to conclude that the most important long-term effect of introduced diseases was some general deterioration in health. Other factors contributed towards increasing morbidity, such as the adoption of Western clothing and the clustering of houses into more compact nucleated villages under mission and, later, Government instruction. Traders and missionaries combined to promote “the fanatic dogma of clothes” (as Grimble later described it) and there was a resulting decline in use of coconut oil for body protection. Without education in hygiene relevant to changes in styles of dress and density of housing, the outcome was all too often dirty, damp clothing, insanitary living conditions and tubercular infection. The latter ailment became one of the most important causes of mortality in the area during the early 20th century. Where diseases caused high (by local standards) mortality on particular islands there was undoubtedly significant decline in numbers. But, as has been demonstrated for other parts of the Pacific with more complete statistics relating to this issue, recovery was probably fairly rapid (see, for example, McArthur 1967). Populations in Kiribati and Tuvalu were not “closed” in the sense that growth and survival depended on the balance between fertility and mortality levels. As noted earlier, there were strong customary links between groups on different islands maintained by regular visiting and intermarriage. While these extra-island ties fostered some dissemination of contagious diseases, they were also very important in the context of demographic recovery after a catastrophic epidemic.

A more important cause of depopulation in certain islands in Kiribati was continuing inter-group warfare. Firearms, a European innovation, were of special significance in the decimation of populations on Kuria and Aranuka in the early 1860s by Tem Baiteke, uea ‘high chief’ of Abemama. However, in - 211 another battle which is believed to have caused considerable mortality (the “religious” war on Tabiteuea in 1880), oral tradition is emphatic that the only European weapon involved was an old cannon in which the vanquished had placed their trust. The role of firearms as a cause of higher mortality levels in local wars is difficult to determine; the psychological advantage obtained through mere possession of guns was probably much more important in a fight than the effect of the bullets (Shineberg 1971; Howe 1974). Maude (1970: 210-1) argues that it was a monopoly over firearms that permitted Tem Baiteke to reduce the populations of Kuria and Aranuka to about 100 apiece from Randell's estimated totals in 1861 of 1,500 and 1,000 respectively. But, while the traditional defensive armour was completely ineffective against bullets, it is important to note that death came to many Kurians and Aranukans at sea after they had fled their islands in canoes. Bingham, a resident missionary on Abaiang at the time, noted that most of those who left Kuria perished at sea because only two canoes were reported as reaching the neighbouring island of Maiana (Maude 1970:211). The fugitives from Aranuka seem to have been more fortunate: Bingham mentions that the populations of Tarawa, Maiana and Abaiang were “somewhat increased” by arrivals from this island. In the case of the war on Tabiteuea in 1880, which was instigated by the “Christians” of the north seeking to punish their southern kin for rejecting “the light” in favour of paganism, the population in the south of the atoll was reduced considerably. Deaths here were estimated variously at between 600 and 1,500; the north Tabiteueans had caused this carnage using traditional weapons. 26 A significant factor underlying decline in population in parts of Kiribati, then, was increasing conflict and disharmony within island communities accompanying diffusion of Western religions, alcohol, firearms, and a demand for coconut oil in an environment where surplus production was difficult in some areas without prejudice to the subsistence system.

Fertility

As far as fertility levels are concerned it appears the early period of prolonged contact with the West saw a rise in birth rates in Tuvalu and possibly in southern Kiribati. If one can accept occasional references to increasing numbers and proportions of children in reports for Tuvalu from the 1870s, it would seem that family sizes were increasing. 27 Resident Samoan pastors and their itinerant European superiors certainly acted quickly to abolish some traditional constraints on population growth. Infanticide and abortion were banned and, although mission rulings were not strictly adhered to, both were practised on a much reduced scale from the 1870s. Polygyny was also forbidden and “men with more than one wife were asked to choose their favourite and send the others back to their natal groups” (Brady 1972:17). But illegitimacy, regarded with disfavour by the Samoan pastors, was con- - 212 sidered to be more acceptable than abortion. There was a demand among childless couples and the elderly for children, and adoption of unwanted babies as well as nieces and nephews was a long-established practice (Kennedy 1931:264).

A number of writers have commented on the speed with which Samoan pastors were able to transform aspects of Tuvalu social organisation and behaviour (Kennedy 1931; Brady 1975; Macdonald 1971; Munro 1979). Kennedy (1931:295), for example, has stated that by the early 20th century a great number of ceremonies associated with ancient rites and customs, games, courtship, marriage, birth, adult initiation, and death had been abandoned. Brady (1975) believes that long-term compatibility with Samoan culture contributed significantly to acceptance of Samoan missionaries, and thereby to rapid acceptance of Christianity. The Samoan pastors demanded, and were eventually given, respect, authority and badges of rank previously accorded to high-ranking chiefs and, from this position of strength in the political system, they made every effort to remove all that pertained to the poo uliuli ‘days of darkness’.

The net demographic effect of mission influence in Tuvalu seems to have been a rise in birth rates. Steady growth in the group's population between 1870 and 1900 lends support to this hypothesis (Table 2). However, north of Beru in Kiribati there is very little evidence that birth rates rose following mission intrusion. Here early mission agents were Hawaiians—“outsiders” as far as the Micronesians were concerned. 28 When Grimble wrote most of his ethnographic essays during the 1920s he had no difficulty obtaining information (and contemporary evidence) of many traditional customs. He did not find, as Kennedy (1931:295) claims was the case in Tuvalu, a culture so altered by Christianity “that even the most tedious and painstaking enquiries fail to reveal more than a few of the most salient features” of the pre-contact social system. Indeed, it was the colonial administration, rather than missions, which eventually curbed many customary practices associated with birth, initiation, marriage and death in Kiribati. In this regard Macdonald's (1972) comments about change on Nonouti are probably relevant to most larger islands in Kiribati. He noted that before 1900 the structure of society and island government changed very little, and innovations were always accepted in such a way that they might be blended with traditional practices. Although most Nonouti residents, were nominal Catholics or Protestants by 1892, neither mission was able to dominate the island's government in the manner which had become characteristic in smaller islands to the south.

A number of customs such as wife-exchange and polygyny were at first discouraged and then forbidden under penalty of heavy fines. But in northern and central Kiribati missionaries met with limited success in their attempts to regulate what they termed “barbaric” or “heathenish” practices, especially - 213 before the establishment of effective British rule. Infanticide certainly declined, but abortion was practised well into the 20th century. It was made illegal under Native Regulations in the colonial period but there were no prosecutions; such a matter was considered to be personal and not public, and evidence was virtually impossible to obtain (Maude 1938:4). Finally, it should be noted that the population for Kiribati as a whole appears to have declined between 1850 and 1900 rather than increased as in Tuvalu. The reason for this was not so much variations in the impact of disease mortality, or changing birth rates; rather it is to be found in very different histories of overseas labour migration before and during the early years of the Protectorate.

Overseas Labour Migration

Between 1850 and 1900 approximately 11,000 Micronesians and 500 Polynesians were transported from these islands to other parts of the Pacific, Australia, Central and South America. Although there is, as yet, no detailed history of 19th century overseas labour migration from Kiribati and Tuvalu, mention is sometimes made of the “devasting effect of this recruitment” on population numbers (Pusinelli 1947:1). Closer inspection of the available evidence suggests that the impact of such migration on population change differed throughout the two groups, as well as over time. Although a number of Micronesians had earlier worked on whalers, the first formal recruiting took place in 1847 when 22 men were taken from Arorae and Tamana to Australia to work on Benjamin Boyd's sheep and cattle stations in New South Wales. The venture proved a failure as most of the labourers deserted from Boyd's employment, found their way to Sydney, and eventually were repatriated. 29 In the 1850s French recruiters were active; some from Kiribati were taken to Reunion, others probably to New Caledonia (D. Shineberg personal communication to B. K. Macdonald). There was a lull in recruiting until 1863 when 161 Micronesians, including 50 from Onotoa, 50 from Arorae and 25 from Nonouti, were kidnapped and taken to Peru. The labour vessel Ellen Elizabeth arrived in Peru after licences to recruit and employ Pacific Islanders had been withdrawn and, when refused permission to disembark passengers, the 110 survivors were taken to Penrhyn Island in the northern Cooks where they were obliged to leave the ship. Some were subsequently hired to work in Tahiti, the first people from Kiribati to do so (Maude and Leeson 1968:268; H. E. Maude personal communication to B. K. Macdonald).

Raids by Peruvian recruiters initiated overseas labour migration in Tuvalu. In 1863 several ships called seeking labour but only Nukulaelae and Funafuti lost significant proportions of their residents through kidnapping. At Nukulaelae about 250 from an estimated population of 300 were enticed on board two Peruvian vessels with the promise of being taken to a nearby island for six months where they could make coconut oil and learn Christianity, before - 214 being returned home with good payment and mission teachers. At Funafuti 171 from a total of about 300 were similarly inveigled aboard (H. E. Maude personal communication to B. K. Macdonald; see also Newton 1967; Murray 1876; Whitmee 1871). The demographic implications of this forced migration were profound for both islands; Nukulaelae lost 80 percent of its population and Funafuti 60 percent in one year. When Whitmee visited Nukulaelae in 1870 he found that two-thirds of the adult population were women, while at Funafuti in the following year, Powell observed that women greatly outnumbered men and children slightly outnumbered the women. Population recovery appears to have been fairly rapid, however. Although there were only 22 adult males on Funafuti in 1873, by 1883 there was an “enormous” number of children, and in 1892 the population numbered 230 — only 70 less than the 300 estimated to be on the island before the Peruvian raids (Davies 1873, SSL 34/2/D; Goward 1892, SSR 2/142). On Nukulaelae replacement was somewhat slower; the island was not visited so frequently by local or foreign shipping and therefore attracted fewer immigrants. The total population in 1892 was stated to be 95, just less than one-third of its pre-Peruvian estimate (Davies 1892, HMS Royalist, RNAS 17).

Following this episode there was little recruiting in Tuvalu. Samoan pastors were strongly opposed to labour migration and this, coupled with small resident populations, made islands here an unpromising recruiting field compared with some islands in Kiribati and the main sources of plantation labour in the western Pacific. Between 1870 and 1900 some Polynesians were recruited for work in Fiji, Hawaii, Samoa and, in the mid-1890s, for Australia. But overseas labour migration was not a significant variable affecting the course of population change in Tuvalu's communities after 1863.

In Kiribati, however, recruiting for employment in Fiji (1864–1895), Tahiti (1867–1885), Samoa (1867–1895), Hawaii (1877–1887), Mexico (1891), Guatemala (1892) and Queensland (1895) ensured that there was a fairly consistent circular movement of people away from and back to their islands. The largest number of migrants went to Fiji — initially there were kidnappings, but it is clear from the numbers that recruited during the 1870s that many were willing to work outside the group. Between January 1870 and December 1875 at least 27 recruiting vessels visited the area, and over 900 Micronesians signed on for terms of three to five years in Fiji. Over the next 20 years a further 35 voyages obtained more than 1,000 recruits. Thus, there were over 1900 known adult migrants between 1870 and 1895. Indications are that recruiting between 1864 and 1870 would account for a further 1,000 adults. 30 The high point in this migration coincides with drought conditions, especially in southern Kiribati. In 1878 and 1883, particularly dry years, over 200 were taken, mainly from Beru and Tabiteuea. The only other year in which adult migration to Fiji exceeded 100 was in 1891 when more than 50 left Tabiteuea. - 215 The main sources of recruits between 1875 and 1895 were Beru and Tabiteuea, where over 350 and 200 respectively are known to have worked in Fiji. Nikunau, Onotoa and Arorae provided between 100 and 150 over the same period while Abaiang, Maiana, Abemama, Nonouti and Tamana provided fewer than 100 each.

Recruiting was not always popular in the island communities. Angry relatives occasionally dragged potential recruits out of the boats by their hair and administered severe beatings to discourage future attempts. To avoid such expressions of displeasure men often swam to recruiting vessels at night, and at least one group at Nikunau took advantage of their fellows' Christian fervour and escaped to a vessel during Sunday service. 31 Not all destinations were as popular as Fiji. The Micronesians quickly learned to fear the “menstealing ships” from Tahiti which began seeking recruits in Kiribati in 1867. Between 1867 and 1872 a total of 687, mostly from the southern islands, were taken by men who “had great sport in the bush catching them and making them fast” (Oates 1871, quoted in Maude and Leeson 1968:269). The Micronesians did not remain entirely passive when subjected to this kind of treatment. In 1870 Whitmee (1871:29-31) was greeted at Arorae and Tamana with knives, hatchets and “every kind of weapon they could find” until his purpose became known. The year before, 80 men from Kiribati, who had been transferred from the Anna (out of Fiji) to the barque Moaroa (destined for Tahiti), broke clear of the hold, killed the captain and some members of his crew, and then escaped over the side. Most reached shore safely at Nikunau (Newbury 1956:164). There was no further recruiting from Kiribati for Tahiti until 1885 when the Forçade de la Roquette obtained 165 labourers (90 men, 75 women) and 46 children. About half were conscripts from Nonouti “sold” to the recruiters by Tem Binoka of Abemama. The majority were subsequently repatriated in 1887 (Newbury 1956:307). It should be noted that many of those who were taken to Tahiti never returned; neither the Tahitian Government nor the planters adhered to repatriation undertakings, mainly because of the cost involved in charters to an area not served by regular shipping.

From 1867 to 1880 Kiribati was the major recruiting area for German planters in Samoa. Over 1,500 people from Kiribati were recruited during the 1870s, the vast majority of whom worked on the plantations of J. C. Godeffroy & Son (Firth 1973; Moses 1973). From 1881 to 1884, however, German recruiting in Kiribati almost ceased as a result of Godeffroys' decision to seek plantation workers in Melanesia. During the mid-1880s, following a reorganisation of Godeffroys' overall trading and plantation strategy, recruiting in Kiribati resumed and the group became a significant back-up to labour supplies from elsewhere. For this reason, Germany pressed the British Government to declare a Protectorate over the area (which fell within Britain's sphere of interest as defined by the Anglo-German Agreement of 1886) and - 216 prevent encroachment by Americans in what they regarded as a “traditional” recruiting ground. But in the decade after declaration of the Protectorate only about 200 Micronesians were indentured for employment in Samoa. Labour recruitment was subject to much closer regulation under the new administration and the Germans found that “without the whip, the imprisonment and irons they cannot get the same work out of the Gilbert Islanders as before”. 32 On the basis of Firth's (1973) detailed analysis of German recruitment and employment of Pacific Islanders, it would seem that about 2,500 Kiribati adults, together with a small number from Tuvalu, worked in Samoa between 1867 and 1895.

Migration from the Pacific Islands to Hawaii between 1877 and 1887 has been reviewed by Bennett (1976). Kiribati was the main source of potential immigrants, and in the decade referred to above, some 1,500 adults and 300 children went to Hawaii (calculated from Bennett 1976:26-7). Although a prime intention in seeking labour in the south Pacific was to stimulate recovery of a declining Hawaiian population, most Micronesians wanted to return home after their three-year contracts terminated. Conditions of employment and wages were decidedly superior in Hawaii, compared with Fiji and Australia, but “nostalgia, satisfied curiosity and a disenchantment with working on strangers' plantations caused the vast majority to take advantage of the promised return passage” (Bennett 1976:22). Hawaiian missionaries in Kiribati were generally opposed to emigration, but in times of hardship, especially after wars or famine, little difficulty was experienced in obtaining willing labourers. Unfortunately, few details are available on islands of origin of recruits; only totals and numbers of men, women and children were registered. Loomis (1970:343-7) has argued that over 1,500 Micronesians were stranded in Hawaii in 1881 “when repatriation came to an end”. She stressed that it was only through efforts of missionaries such as Bingham that some of these people were finally given passage back to Kiribati in 1903. Loomis figure of 1,500 is probably too high, and Bennett (1976:26-7) has argued that between 1881 and 1887 Government vessels and private charters continued to visit Kiribati regularly, and a significant number of Micronesians were returned to their home islands. No doubt many were lost to Hawaii's graveyards and Honolulu's slums (Lambert 1975:225), but it would seem that most of the survivors eventually returned to the group.

The final episodes in long-distance overseas labour migration involved two short-lived movements of Micronesians during the 1890s. Between 1890 and 1892 about 1,000 adults recruited for work in Central America. Three ships were involved — the Helen W. Almy in 1890 (300 recruits), the Tahiti in 1891 (300 recruits, 100 children) and the Montserrat in 1892 (400 recruits). The Tahiti, bound for Mexico, capsized with total loss of life off the southern California coast. The Montserrat took its recruits to Guatemala. Although we - 217 have been unable to ascertain the specific destination of the Helen W. Almy, it is known that, in 1896, this vessel repatriated 203 adults from its own recruiting voyage and that of the Montserrat. Of the remaining recruits, 39 adults were re-engaged, 68 were reported as dead, and 5 had settled in Mexico. The fate of the remaining 400 is unknown. 33

The second short-lived movement during the 1890s was to Australia. Although a few migrants made their own way to Queensland during the 1880s and early 1890s, formal recruiting for this area from Kiribati and Tuvalu did not commence until 1895. In that year the May recruited 62 males and 11 females, and the Lochiel 104 males and 13 females — a total of 190. Of this numbers, 165 were from Kiribati (46 from Beru and 23 from Tabiteuea) and within five years 147 had returned. 34 The Queensland trade was stopped after these two voyages when the High Commissioner, Sir John Thurston, protested on somewhat vague grounds that service in Australia would be detrimental to the interests of the Micronesians and politically embarrassing to the new Protectorate administration. Thurston preferred these islanders to work in Fiji.

By the end of the century there were beginnings of further labour migrations which were to prove crucial to the economic future of Kiribati and Tuvalu. From the 1880s about 100 Micronesians were employed on Fanning and Washington, islands that were later to become sites of Government-owned plantations in the Gilbert and Ellice Islands Colony. And in 1900, with Albert Ellis' discovery of phosphate on Ocean Island, there began a labour migration that was later extended to Nauru, and was to involve thousands of islanders from the two groups over the next 75 years.

Between 1847 and 1895 approximately 9,300 adults from Kiribati were recruited for work in areas outside the central Pacific. A summary of the volume of movement by destination is given in Table 3. It should be noted that these figures do not include one important component in the flows away from and back to Kiribati — children who were too young to be legally classed as “recruits”. An important feature of overseas labour migration from these islands in the 19th century was the involvement of entire families. On 24 of the 35 voyages to Fiji between 1876 and 1895, for example, there were about as many women and children (475) on board as there were adult males (481) (Table 4). Incomplete records of the sex and adult/child status of those who went to Hawaii, Tahiti and Central America indicate that this family migration was widespread. Indeed, of 2,662 migrants covered in those voyages for which we have the relevant data, less than half (49 percent) were adult males; 35 percent were women and 16 percent children (Table 4). The ratio of children to adults varied considerably, but it would seem that one child per five or six adults could have moved from Kiribati to the main employment centres during the second half of the 19th century.

- 218

Labour Migration and Depopulation

Although mortality rates were higher on ships and at employment sites than in Kiribati during a “normal” year, the labour trade did not “cause a great diminution of the population” (Pusinelli 1947:2) before establishment of the Protectorate. There was certainly a net loss in population to small islands, but most of those who left the group to work overseas eventually returned to Kiribati. Of the 1,119 adults who went to Fiji between 1876 and 1895, 140 (or 12.5 percent) have their deaths recorded in the Register of Polynesian Immigrants. Male mortality was slightly higher than that of females — probably reflecting differences in type of employment and living - 219 conditions. These figures can be questioned on several grounds, the most obvious of which is that known repatriations (629) plus registered deaths account for less than 70 percent of the migrants. But there is other evidence to show that Micronesian mortality in Fiji was generally lower than that for the migrant population as a whole. For example, between 1878 and 1882 when sugar plantations were being established, mortality among all immigrants in the first year of indenture was 14.6 percent. The proportion of Micronesians who died was much lower (4.8 percent) mainly because they were kept out of the sugar areas and employed on coconut plantations. 35 In Hawaii, Bennett (1976:21-2) reports that mortality among Micronesians from Kiribati was higher than for Melanesians from the New Hebrides; in this regard it is interesting to note that the former were employed almost exclusively in the sugar industry with which was associated high levels of mortality among Pacific Islanders in Fiji.

There were other factors contributing towards lower than average mortality among Micronesians in Fiji besides work outside the sugar areas. The diet seems to have presented few problems and as most were engaged in familiar work on coconut plantations, adjustment to the new life was not too difficult. But there is another factor which may have contributed towards general good health and contentment — the large number of Micronesian women and children who were taken to Fiji (Table 4). Corris (1973:45-6) has demonstrated that labour recruiting in the western Pacific was dominated almost entirely by men; official returns of New Hebrideans and Solomon Islanders in Queensland and Fiji, for example, show that between 5 and 10 percent of all recruits were women. As noted earlier, the sex ratio of Micronesian migrants was much more balanced; indeed, for some areas more women than adult men are listed in recruiting journals. There are several reasons for this, not the least important of which is that plantation employers welcomed women and children because they could be employed at lower wages than had to be paid to men. In many plantation jobs work output from women and children was just about as high as that from adult males.

In Kiribati female migration was favoured by a number of social and cultural factors: the younger age at marriage for men in the southern islands which accompanied mission pressure intended to reduce sexual promiscuity; the jealousy with which men regarded their wives possibly prompted many recruits to take their spouses with them; local community leaders in some islands only permitted recruiting among the married “provided both man and wife went together”, 36 and the keen demand for nikiranroro in “domestic service” (a term used to describe a wide variety of jobs in Fiji). Nikiranroro were free agents when it came to recruiting. Unlike the situation in the New Hebrides and Solomon Islands, where all women were regarded as valuable assets and whose recruitment hence was generally forbidden, there was little

- 220

- 221

in community attitudes to encourage nikiranroro to stay, especially after mission teachers had reacted against the tolerance with which they were traditionally regarded.

It would seem that recruiting did not result necessarily in mortality levels that were always substantially different from those in the country of origin. Bennett (1976:22) mentions Hawaiian missionaries in Kiribati who, while opposed to labour migration, still held the opinion that “the percentage of deaths from violence in the group is greater than the percentage from all causes among Gilbert Islanders in the Hawaiian Islands”. However, labour migration was undoubtedly an important cause of temporary depopulation on certain small islands during the latter half of the 19th century. In addition to the cases of Nukulaelae and Funafuti mentioned with regard to the Peru slave trade, Tamana, Arorae and Beru had significant proportions of their populations absent at different times. For its size Arorae was probably more affected by recruiting than any other island in southern Kiribati. As we show in the next section, some estimates of the population on this island in the 1870s were as low as 400, while in the 1880s and 1890s estimates in the range of 900 to 1,200 are more common. This fluctuation is consistent with reported high mortality during a drought in the early 1870s combined with heavy recruiting for Mexico, Tahiti, Samoa and Fiji, followed in the 1880s by repatriation, a much lower incidence of recruiting, intensification of mission activity, and a rise in the rate of natural increase.

To sum up, the effects of overseas labour migration on the nature of population change were not the same throughout Kiribati. Normally the number of persons leaving a given island in any one year did not exceed 100. Because almost as many women as men left to work elsewhere, the sex ratios of island populations were not distorted to the extent they were in other parts of the Pacific. Age composition was certainly affected but in most cases the majority of adults were only temporarily absent. The loss to potential population increase caused by labour migration was not nearly as great as has sometimes been claimed.

Analysis of factors influencing the three demographic processes has led us to the conclusion that Kiribati and Tuvalu did not suffer the great population decline suggested when Tupper's figures for 1901 are compared with Randell's and Becke's estimates. Becke's precontact population of 20,000 for Tuvalu is a gross exaggeration. There is also good reason to believe that most of Randell's often-cited estimates for Kiribati about 1860 are far too high. Even though labour recruiting was well established by the time more reliable estimates were recorded in the 1870s, it is difficult to accept that populations on - 222 some of the southern Kiribati islands had declined by as much as 80 percent in a decade.

In this final section we suggest, and attempt to justify, a more “realistic” set of estimates for island residents in the early 1860s. Although we have a record of numbers resident on most islands in Kiribati in 1852, we have chosen to concentrate attention on Randell's estimates published in The Friend in September 1861 because these are the figures most frequently cited for early contact populations in Kiribati. They also correspond more closely in time to the first mission estimates of island populations in Tuvalu. In our assessment of population estimates for Kiribati we have divided the group into three areas: the northern islands (Makin to Tarawa), the central islands (Maiana to Abemama), and the southern islands (Nonouti to Arorae). While all islands appear to have experienced an overall decline in numbers in the latter half of the 19th century, the magnitude of this “depopulation”, along with the reliability of Randell's estimates, varies considerably in the three areas. Numbers and trends in Tuvalu are then briefly discussed in the context of Newton's (1967) appraisal of mission estimates. (Tables containing details of the estimates are grouped at the end of the text.)

Northern Kiribati (Tables 5 and 6)

The most complete record we have for islands in the northern group is for Makin and Butaritari where estimates made by traders and visitors can be supplemented by those given by early Hawaiian missionaries. Except for Hale's (1846:93) high estimate of 5,000 for the two islands in 1841 (a figure he obtained from a British seaman, Robert Grey [alias Wood], who had deserted his ship and taken up residence in the area in the late 1830s), most populations cited for Makin and Butaritari before 1900 are between 2,000 and 3,000. Randell was resident almost continuously on Butaritari from 1846 to 1870 and, because trading brought him into regular contact with people from both islands, we consider his combined figure of 2,000 in 1861 is a reasonable estimate. That the population changed little over the latter half of the 19th century is also not unreasonable given that these islands were never a major recruiting centre, nor was the population reported to have been seriously affected by alien diseases. 37 When Randell's estimate of 2,000 is compared with an adjusted population for 1901 (Tupper's figure of 1,712 does not appear to represent a complete enumeration for the two islands), the total decline recorded for the 40 years is 5 percent. 38 A comparatively low crude density of 99 persons per square kilometre is not unexpected given the number of uninhabited islets scattered around the periphery of Butaritari's large lagoon.

In the case of Marakei, both Randell and Davis make estimates of 2,000 while in 1871 The Friend mentions a population of 1,000 and Tupper returned - 223 a total just under 1,500 for 1901. Some net loss in numbers may have occurred as a result of sporadic warfare through the latter part of the century, as well as some recruiting for Fiji, Hawaii and Guatemala. However, an overall decline of 25 percent between 1860 and 1900 seems excessive; there is no evidence to suggest such heavy depopulation over the period. As the 1901 population figure for Marakei is not unduly low compared with other enumerations for this island about the turn of the century, we therefore consider Randell's estimate of 2,000 is rather high. A figure in the region of 1,600 and a decline over 40 years of between 6 and 10 percent might be more reasonable.

Much more severe inter-group fighting and labour migration as a means of escaping this were common on Abaiang and Tarawa. An overall decline in numbers is highly probable here, although it should be recalled that both islands did receive immigrants from Aranuka in the early 1860s. Gulick (1852), Randell (1861) and The Friend (1871) cite estimates of 3,000-3,500 for Abaiang and Tarawa, while figures ranging from 2,000 to 2,800 are mentioned in early administration reports. In light of political instability and recruiting, population decreases by 15 to 20 percent between 1860 and 1900 are probably not unrealistic. This would allow for a population of about 2,700 for Abaiang and one of 3,000 for Tarawa in 1860, slightly lower than Randell's estimates. Greater population densities for Marakei, Abaiang and Tarawa, compared with the average for the two northern islands, are acceptable given that settlement was spread over more of their total land areas.

Central Kiribati (Tables 7 and 8)

For the four central islands — Maiana, Abemama, Kuria and Aranuka — our record of estimates before 1900 is the least satisfactory. Early estimates by Gulick (1852) and in The Friend put the combined population for these islands between 9,500 and 14,500. Populations between 3,000 and 5,000 for Maiana and Abemama seem much too high, however. Although intermittent warfare on Maiana may have contributed to some decrease in numbers, this island was visited rarely by recruiting vessels, perhaps because of navigational difficulties posed by a shallow lagoon. There is thus little evidence to support a population decline of the magnitude suggested by a comparison of Randell's estimate with the 1901 census figure. Indeed, Maiana's population was supplemented from time to time during the 1860s by refugees fleeing the rampages of Baiteke and Binoka of Abemama (Maude, 1970). In 1881 a German trader told Maxwell (on HMS Emerald) that there were about 1,600 living on the island and this is a figure not too dissimilar to Rooke's estimate in 1886 of 1,700. A more reasonable estimate of the population in 1860 would seem to be about 1,800.

Abemama, Kuria and Aranuka present a number of problems because of - 224 difficulties of assessing the impact of wars initiated by Baiteke and Binoka. Randell's estimate for the three islands was 7,500; The Friend in 1871 gave a total of 6,500. The major period of inter-island hostilities started in 1863 with Baiteke's attacks on the other two islands. In addition to the deaths and emigration caused by the fighting, 39 there is evidence to suggest that the populations of Kuria and Aranuka were kept at about 100 to 150 each during the latter part of the 19th century (Maude 1970:211). Most residents were in a state of servitude under Abemama with, as Maude notes, “the increase being shipped periodically off to Abemama”. It would seem likely, therefore, that of the three islands, Abemama suffered the least decline in population. Prewar estimates of 1,000 each for Kuria and Aranuka are probably too high, but numbers of this magnitude give low population densities which are consistent with traditions for these islands. It might also be noted that massive depopulation is not suggested in the oral traditions of these wars.

Maude (1970) has described at length events during the reigns of Abemama's two most powerful uea ‘high chiefs’. Baiteke and Binoka apparently succeeded in restricting labour recruiting from their domains and, rather than losing many people through this cause, Maude (1970:218) argues that “with surplus land and a dwindling population, immigrant labour was encouraged to settle on Abemama”. Although there was little labour migration overseas, the population of Abemama is reputed to have decreased during the late 19th century. Maude (1970:216) points out that for some unknown reason many women did not have children. Mention of the absence of children was made on occasion by visitors (e.g., Woodford in 1884) and adoption from neighbouring islands was common. When the Lands Commission attempted to settle claims on Abemama in 1948 it was found that disputes most commonly arose when people died without issue. Of the 1,416 Abemamans whose deaths were registered between 1898 and 1948 the Lands Commission found that almost 70 percent (973) had died childless. 40

Although we do not have any precise information on unusually high levels of infertility on Abemama during the 19th century, an 80 percent drop in population between 1860 and 1900 is unlikely, if only because there was periodic immigration from Kuria, Aranuka and Nonouti. While we are obviously only guessing, we suggest a more realistic figure for the island about 1860 would be in the region of 2,500 inhabitants — half that suggested by Randell. Such an estimate would still allow for a 55 percent decline over 40 years. Even though Randell spent many years in Kiribati, and frequently visited Abemama on trading voyages, like all foreigners he was restricted to landing at only one place — the little Entrance Islet on the western reef (Maude 1970:259). His estimates of Abemama's population were not based on the familiarity with place and people he had in Butaritari and Makin.

- 225

Southern Kiribati (Tables 9 and 10)

Randell's estimates for most islands in southern Kiribati seem quite unrealistic in the light of figures cited for other years in the late 1860s and early 1870s. The most detailed set of guesses and counts which we have for Kiribati relate to five islands which became the London Missionary Society's domain in the group. These were also the islands which experienced the earliest and perhaps most continuous labour recruiting. In spite of extensive migration overseas, populations in the south did not suffer such severe declines as some in central Kiribati. It was also in this area that missionaries made most rapid progress in quelling practices such as infanticide and abortion. Indeed, on certain islands there appears to have been quite considerable population increase between 1880 and 1900. Although there are anomalies in the data and the problem of determining whether de facto or de jure populations are mentioned cannot be resolved, some fluctuation in numbers is consistent with recruiting trends, environmental hazards, and disease mortality reported for the 30 years after 1870.

Nonouti and Tabiteuea are the two largest islands in Kiribati and, in the 19th century, contact tended to be limited to the more accessible villages. Randell had resident agents on both islands in the mid-1860s but, as Maude and Leeson (1968:260) note, “by and large the trader was dependent for trade and personal safety on his relations with the people of the village where he set up his headquarters, even though in times of peace the whole island might trade with him on sufferance.” The populations for these two places in 1867 and 1871, which are cited in The Friend, are based on assessments by traders, and it is interesting to note that their figures are much lower than Randell's suggestions.

For Nonouti estimates between 1867 and 1900 all fall within a range from 2,500 to 3,000. Labour recruiting from 1863, Tem Binoka's depredations in the early 1880s, 41 migration to Abemama, and some effects of “an environment of war, murder, drunkenness and theft” (Maude and Leeson 1968:273), certainly resulted in a decrease in population. It seems unlikely, however, that this decline was of the magnitude of 50 percent which emerges from a comparison of Randell's lowest estimate with Tupper's figure in 1901. A more reasonable population estimate could be one which is slightly higher than that given in The Friend in 1871; we suggest a range from 3,400 to 3,700 for Nonouti in 1860. Tabiteuea is credited with a population of between 5,000 and 6,000 in the years before the civil war of 1881, an outbreak of measles in the 1890s, and significant levels of labour migration over the latter part of the century. Given estimates of about 4,000 in the 1890s, a population in 1860 of between 5,000 and 5,500 seems realistic. This would allow for a net decline of between 18 and 23 percent over the subsequent 40 years.

Although islands from Beru south are more prone to drought than any - 226 others in Kiribati, some of the highest population densities are reported here. Beru and Tamana are known to have had quite dense populations from the earliest mission contact. However, densities of the level suggested by Randell's estimates were most improbable for all islands with the exception of Beru, where Randell has a rather low range for the island's population. Beru was a major labour recruiting area throughout the period 1865 to 1895 and, even though the island had Samoan pastors in residence and was not beset by civil war, a pre-recruiting population which is slightly larger than the one recorded for 1901 seems likely. Estimates in the late 1860s and early 1870s set the population about 3,000 but we would suggest a lower figure within the range 2,300 to 2,500. A drop in numbers of 400 or more in the 1870s, when there was a serious drought and heavy recruiting, also fits with this estimate.

Nikunau has never had a population the size of that on Beru since reliable counts were made, even though the island is a larger one. Estimates in the 1880s, after the mission was well established, range from 1,671 (1881) to 1,912 (1885). The island was visited fairly regularly by recruiting vessels, but it is not so badly affected by drought as others in southern Kiribati. Its population may not have experienced very heavy labour migration or drought and disease mortality. A pre-recruitment estimate of about 2,000 seems more feasible than various figures between 3,000 and 6,000 cited for the 1860s and early 1870s. This still allows for an overall decrease by 16 percent on the 1901 population reported by Tupper. The atoll of Onotoa, on the other hand, is particularly susceptible to drought and was a major source of recruits. An early resident trader (Waters) working for Randell on Onotoa in the late 1860s mentioned a population of between 1,200 and 1,400 in 1867. In the early 1870s figures between 1,500 and 1,700 were reported and it would seem that some 1,600 persons may have been resident on the island in 1860.

On the basis of figures contained in Table 9 a case can be made for significant depopulation in the 1860s and 1870s on the small drought-prone reef-islands of Tamana and Arorae. Both islands were visited by recruiting vessels from the early 1860s and a serious drought in the 1870s is known to have caused extensive mortality as well as heavy recruiting for Tahiti, Samoa and Fiji. On Tamana, for example, Maxwell (1881, HMS Emerald) mentions a trader's estimate of about 800 “carried off” by drought. A clearer indication of the relative contribution of mortality and migration to population decline in Tamana is given by Turner (1878, SSL 173). He records that during 1877 there were 216 deaths due to ‘starvation’ and 121 departures for overseas destinations leaving a population early in 1878 of 282. Numbers on these two small islands fluctuated considerably during the 1870s and early 1880s as a result of labour recruiting and repatriation. While population estimates for Tamana and Arorae vary markedly between 1867 and 1871, it seems unlikely that the former had about 3,000 residents in 1861 and the latter between - 227 2,000 and 2,500 as is suggested by Randell. It is difficult to account for declines of 100 and 200 percent between 1861 and 1871, even though the islands were sources of labour for Fiji, Samoa and Tahiti. More realistic estimates for the early 1860s would seem to lie in the ranges 900 to 1,100 for Tamana and 1,200 to 1,400 for Arorae. Such totals still allow for quite significant overall population decline before 1900.

Kiribati Group: A Summary (Table 11)

Our estimates of island populations in 1860 produce totals of 31,500 and 35,100 for Kiribati; 19,000 less than Randell's range of 50,500 to 54,000. Differences between the two sets of estimates tend to increase with distance from Butaritari, which is not unexpected given that Randell relied to a large extent on his trading agents for information on islands to the south. When our group totals are compared with a population for 1900, the overall net losses in numbers during the 40 years range from 5,650 to 9,250, representing 18 and 26 percent respectively of the lower and upper estimates for 1860. The central islands emerge as the group where decline was most severe; those to the north have slightly lower rates of decrease than the seven southern islands.

This pattern of population change is reasonably consistent with major events known to have affected demographic development in different parts of Kiribati. Of the 11,000 or so who are known to have gone overseas, at least half returned at some stage before 1900. Thus the net loss from labour migration probably did not exceed 5,000 and may have been substantially less. Further, there was exceptional depopulation in the central islands of Kuria, Aranuka and Abemama. Also to be taken into account are the Tabiteuea civil war, Binoka's escapades on Nonouti, the endemic fighting on Tarawa and Abaiang, drought (especially on Tamana and Arorae), the effects of alien diseases, and, on the credit side, additions by birth to the population. These considerations lead us to the conclusion that, on balance, the maximum total net loss to the population of Kiribati over the period 1860 to 1900 probably did not exceed 10,000; in fact it could have been significantly lower than this. Assuming the figures for 1900 are reasonably accurate, this would allow for a population in 1860 of no more than 35,000. In this regard it should be noted that as early as 1873 Bingham was talking in terms of a population of 30,000 for Kiribati, a figure which is much more consistent with numbers about 33,000 in 1860 than the 50,000-54,000 suggested by Randell.42

Tuvalu (Tables 12 and 13)

Gross exaggeration of the pre-recruiting population of Tuvalu has persisted - 228 in official publications, as well as in other writings through to the 1970s (see, for example, Sabatier 1939:136, and Coates 1970:187). The following statement (or similar versions) was published in the Colony's Annual Report for the best part of 40 years:

Newton (1967) has effectively dispelled the myth of a precontact population of 20,000. Although there were occasional outbreaks of measles and dysentery there is no indication that these diseases decimated the population. The only recruiting of any significance was, as noted earlier, Peruvian kidnapping in 1863, and that seriously affected only two islands.

There are numerous estimates and enumerations of island populations from the mid-1860s when the London Missionary Society first sent pastors to Tuvalu. Because the pastors were resident on very small islands (none of which had a population exceeding about 500) and assumed a dominant role in local social and political life, their figures are probably reasonably accurate, especially for the last two decades of the century. While there are some anomalies in figures for certain islands in some years, the general impression one gains about population trends in Tuvalu from the 1880s is of growth. This accords with repeated mention in the London Missionary Society's journals of steady population increases once mission influence was well established.

The earliest complete set of estimates for the group is in 1866 when Murray called at all islands. His figures for Funafuti and Nukulaelae clearly reflect the demographic impact of the Peruvians' visit three years previously. In the light of other estimates, however, those for Nanumea and Niutao seem excessive. Allowing for populations of 300 each on Funafuti and Nukulaelae, and lower numbers on Nanumea and Niutao, a total about 2,800 for the early 1860s seems feasible. This is somewhat lower than Newton's (1967) “early nineteenth century population”, mainly because of a different estimate for residents on Nanumea. When totals for subsequent years are compared with 2,800 for 1860, it is apparent that there was some decline in the populations of all islands before the mid-1870s. In 1876 Turner (1884) claimed there were 2,497 in the group, a drop of almost 11 percent on our 1860 estimate. Nine years later, the population was reported to be just over 2,900 (Newell 1885, SSJ 182) and in another decade it had reached 3,226 according to Swayne (1895) — an increase of 15 percent on the number - 229 believed to be present 35 years previously.

Analysis at the island level reveals considerable fluctuations in numbers. Many of these can be accounted for by inter-island visiting which was then, and still is, an important feature of Tuvaluan life. Two examples from early in the period under discussion demonstrate the problem of comparing figures cited for specific islands. In 1866 Murray (1866, SSJ 157) stated that his population estimate (300) for Nui included “a full 100” from other islands who were either temporary or permanent residents. Whitmee (1871:14) noted that 30 or 40 visitors from Niutao were not included in his estimate of 376 for Vaitupu's population in 1870. There are a mixture of de jure as well as de facto populations in the record summarised in Table 12. For this reason, and because the numbers involved are so small, we have not attempted any analysis of population change for individual islands.

A concluding comment

By 1900 the population of Kiribati and Tuvalu was in the region of 29,400, about 200 less than the figure Davis estimated when he declared Protectorates over the two groups in 1892. Between 1860 and 1900 numbers had dropped by 8,400 at most, from a maximum population of 37,800. Although this represents a decline of some 22 percent, it is not indicative of serious depopulation in all areas, since there were significant variations in population trends in different parts of the two island groups. It is these variations, rather than a generalised assessment of total figures, which must provide the focus for future inquiry into the demographic history of this part of the central Pacific.

Assessment of 19th century population estimates for small islands in the Pacific is fraught with difficulties, and we are very conscious of the fact that our “more realistic” figures are no more than guesses. There is no “hard” evidence on the numbers and composition of populations before the establishment of stations manned by resident missionaries. Because of the nature of the available information we have labelled our study an exercise in “speculation” rather than “re-interpretation”. Obviously all speculations can be challenged, and we expect the estimates cited in this paper to be questioned and be subjected to further reappraisal, rather than be accepted uncritically as the early contact populations of Kiribati and Tuvalu.

We are grateful for the detailed comments and constructive criticism provided by Harry Maude, Eric Bailey and Keith and Anne Chambers on an earlier version of this paper. Needless to say, we remain responsible for any errors of fact and interpretation.

- 230

- 231

- 232

- 233

- 234

- 235

- 236

- 237

- 238

- 239 Page of endnotes

- 240 Page of endnotes

- 241 Page of endnotes

- 242 Page of endnotes

- 243

Unpublished sources

The two most important sources of population estimates for Kiribati and Tuvalu are reports by the commanding officers of Royal Navy ships patrolling in the central Pacific in the 1880s and 1890s, and journals of people associated with the London Missionary Society. The former are identified in the text by captain, name of vessel and archival reference; the latter by missionary and journal number. The naval reports were sighted in records of the Royal Navy, Australia Station which are held in the National Archives of New Zealand. The London Missionary Society journals were sighted on microfilm in the Turnbull Library, Wellington, and the Mitchell Library, Sydney, and most of those referred to in this paper are to be found in the South Seas Journals (SSJ) series. Some have been misplaced among the South Sea Records (SSR) and South Sea Letters (SSL). We refer to the actual location of specific journals in the text. Several miscellaneous documents were also sighted on microfilm in the Mitchell Library — these are referenced by film number and the prefix “ML”.

In addition to these sources, various manuscript and printed series are mentioned in footnotes and tables in the paper. These are:

• ABCFM — American Board of Commissioners for Foreign Missions (consulted on microfilm — National Library of Australia and now also issued by the Pacific Manuscripts Bureau).
• CO 201 — Colonial Office, Original Correspondence, New South Wales series (consulted on microfilm).
• FO 58 — Foreign Office (Great Britain), Original Correspondence, Fiji.
• GEIC 21 — Records of the District Office, Gilbert Islands, 1933–68 (formerly held in the Western Pacific Archives, Suva; now returned to Tarawa).
• QVP — Votes and Proceedings of the Queensland Legislative Council.
• RNAS — Royal Navy Australia Station (held in the National Archives of New Zealand).
• WPHC 4 — Records of the Western Pacific High Commission, Series 4, Inwards Correspondence, General (consulted at the Western Pacific Archives, Suva; now housed in London).

Published sources (including theses)

• BAYLISS-SMITH, T.P., 1974. “Constraints on Population Growth: the Case of the Polynesian Outlier Atolls in the Pre-contact Period.” Human Ecology, 2:259-95.

- 244

• —— 1975. “The Central Polynesian Outlier Populations since European Contact,” in V. Carroll (ed.), Pacific Atoll Populations. Honolulu, University of Hawaii Press.
• BEDFORD, R.D., MS. 1967. Resettlement: Solution to Economic and Social Problems in the Gilbert and Ellice Islands Colony. M.A. thesis in Geography, University of Auckland.
• BENNETT, J.A., 1976. “Immigration, ‘Blackbirding’, Labour Recruiting? The Hawaiian Experience 1877–1887.” Journal of Pacific History, 11:3-27.
• BRADY, I.A., MS. 1972. Population Structure and Distribution in the Ellice Islands. Paper presented at the Pacific Atoll Populations Conference, East-West Center, Honolulu, December 1972. Mimeo.
• —— 1975. “Christians, Pagans, and Government Men: Culture Change in the Ellice Islands,” in I.A. Brady and B.L. Isaacs (eds), A Reader in Culture Change, Vol. 2: Case Studies. New York, Schenkman.
• CARROLL, V., (ed.), 1975. Pacific Atoll Populations. Association for Social Anthropology in Oceania Monograph No. 3. Honolulu, University of Hawaii Press.
• COATES, A., 1970. Western Pacific Islands. London, Her Majesty's Stationery Office.
• CORNEY, B.G., STEWART, J. and THOMSON, B.H., 1896. Report of the Commission Appointed to Inquire into the Decrease of the Native Population. Suva, Government Printer.
• CORRIS, P., 1973. Passage, Port and Plantation: a History of Solomon Islands Labour Migration 1870–1914. Melbourne, Melbourne University Press.
• FINSCH, O., 1893. Ethnologische Erfahrungen und Belegstucke aus der Sudsee (Ethnographic Experiences and Specimens from the South Seas). Vienna, Alfred Holder.
• FIRTH, S.G., MS. 1973. German Recruitment and Employment of Labourers in the Western Pacific before the First World War. D.Phil. thesis in History, Oxford University, Oxford.
• GRIMBLE, A., 1921. “From Birth to Death in the Gilbert Islands.” Journal of the Royal Anthropological Institute, 51:21-54.
• —— 1924. “Canoes of the Gilbert Islands.” Journal of the Royal Anthropological Institute, 54:101-39.
• —— 1931. “Gilbertese Astronomy and Astronomical Observances.” Journal of the Polynesian Society, 40:197-224
• —— 1933–34. The Migrations of a Pandanus People as Traced from a Preliminary Study of Food, Food Traditions and Food Rituals in the Gilbert Islands. Memoirs of the Polynesian Society No. 12. Wellington (originally Supplements to Vols. 42 and 43 of the Journal of the Polynesian Society).
• GRIMBLE, R., 1972. Migrations, Myth and Magic from the Gilbert Islands. Early Writings of Sir Arthur Grimble. London, Routledge and Kegan Paul.
• GULICK L.H., 1862. “Micronesia of the Pacific Ocean.” Nautical Magazine and the Naval Chronicle, 3:358-63, 408-17.

- 245

• HALE, H., 1846. United States Exploring Expedition: Ethnography and Philology. Philadelphia, Lea and Blanchard.
• HEDLEY, C., 1896. “The Ethnology of Funafuti,” in The Atoll of Funafuti. Memoirs of the Australian Museum No. 3, Sydney.
• HILDER, B., 1959. “Polynesian Navigation Stones.” Journal of the Institute of Navigation (London), 12:90-7.
• HORNELL, J., 1936. The Canoes of Polynesia, Fiji and Micronesia. Bernice P. Bishop Museum Special Publication No. 27, Honolulu.
• HOWE, K.R., 1974. “Firearms and Indigenous Warfare: a Case Study.” Journal of Pacific History, 9:21-38.
• — 1978. “Tourists, Sailors and Labourers: A Survey of Early Labour Recruiting in Southern Melanesia.” Journal of Pacific History, 13:22-35.
• KENNEDY, D.G., 1931. Field Notes on the Culture of Vaitupu, Ellice Islands. Memoirs of the Polynesian Society No. 9. New Plymouth, Avery.
• LAMBERT, B., 1975. “Makin and the Outside World,” in V. Carroll (ed.), Pacific Atoll Populations. Honolulu, University of Hawaii Press.
• LE HUNTE, G.R., 1883. Six Letters from the Western Pacific. Colombo (Ceylon), Government Printer.
• LEWIS, D., 1971. “‘Expanding’ the Target of Indigenous Navigation.” Journal of Pacific History, 6:83-95.
• — 1972. We, the Navigators. The Ancient Art of Landfinding in the Pacific. Wellington, A.H. and A.W. Reed.
• LOOMIS, A., 1970. To All People. A History of the Hawaii Conference of the United Church of Christ. Tennessee, Kingsport Press.
• McARTHUR, N., 1967. Island Populations of the Pacific. Canberra, Australian National University Press.
• — 1970. “The Demography of Primitive Populations.” Science, 167:1097-1101.
• MACDONALD, B.K., MS. 1971. Policy and Practice in an Atoll Territory: British Rule in the Gilbert and Ellice Islands, 1892–1970. Ph.D. Thesis in Pacific History, The Australian National University, Canberra.
• — 1972. “Te Reitaki n Nonouti. A Survival of Traditional Authority in the Gilbert Islands.” Journal of Pacific History, 7:137-50.
• McLEAN, R.F. and MUNRO, D., MS. 1976. Nineteenth Century Hurricanes in the Ellice Islands (Tuvalu). Mimeo.
• MAUDE, H.E., 1938. Report on the Colonization of the Phoenix Islands by the Surplus Population of the Gilbert and Ellice Islands. Suva, Government Printer.
• — 1963. The Evolution of the Gilbertese Boti: an Ethnohistorical Interpretation. Memoirs of the Polynesian Society No. 35. Wellington (originally Supplement to Vol. 72 of the Journal of the Polynesian Society).
• — 1970. “Baiteke and Binoka of Abemama: Arbiters of Change in the Gilbert Islands,” in J.W. Davidson and D. Scarr (eds), Pacific Islands Portraits. Wellington, A.W. and A.H. Reed.
• — 1971. “The Cultural Setting,” in Honor Maude, The String

- 246

• Figures of Nauru. Occasional Papers in Asian and Pacific Studies, No. 2. Adelaide, Libraries Board of South Australia.
• —— and LEESON, I., 1968. “The Coconut Oil Trade of the Gilbert Islands,” in H.E. Maude (ed.), Of Islands of Men: Studies in Pacific History. Melbourne, Oxford University Press.
• MOSES, J.A., 1973. “The Coolie Labour Question and German Colonial History in Samoa: 1900–1914.” Journal of Pacific History, 8:101-24.
• MUNRO, D., 1979. “Kirisome and Tema: Samoan Pastors in the Ellice Islands,” in D. Scarr (ed.), More Pacific Islands Portraits. Canberra, Australian National University Press.
• MURRAY A.W., 1876. Forty Years' Mission Work in Polynesia and New Guinea, from 1835 to 1875. London, Snow.
• NEWBURY, C.W., MS. 1956. The Administration of French Oceania, 1842–1906. Ph.D. Thesis in Pacific History, Australian National University, Canberra.
• NEWTON, W.F., 1967. “The Early Population of the Ellice Islands.” Journal of the Polynesian Society, 76:197-204.
• PIRIE, P., 1972. “The Effects of Treponematosis and Gonorrhoea on the Populations of the Pacific Islands.” Human Biology in Oceania, 1:187-206.
• PRICE, C.A. with BAKER, E., 1976. “Origins of Pacific Island Labourers in Queensland, 1863–1904: A Research Note.” Journal of Pacific History, 11:106-21.
• PUSINELLI, F.N.M., 1947. A Report on the Results of the Census of the Population 1947. Gilbert and Ellice Islands Colony. Suva, Government Printer.
• ROBERTS, R.G., 1953. “The Dynasty of Abemama.” Journal of the Polynesian Society, 62:267-76.
• —— 1958. “Te Atu Tuvalu: a Short History of the Ellice Islands.” Journal of the Polynesian Society, 67:394-423.
• SABATIER, E., 1939. Sous l'Équateur du Pacifique: Les Îles Gilbert et la Mission Catholique (On the Equator of the Pacific: The Gilbert Islands and the Catholic Mission). Paris, Editions Dillen.
• SHINEBERG, D., 1971. “Guns and Men in Melanesia.” Journal of Pacific History, 6:61-82.
• THOMSON, B., 1908. The Fijians: a Study in the Decay of Custom. London, Heinemann.
• TURNER, G., 1884. Samoa a Hundred Years Ago and Long Before. London, Macmillan.
• WHITMEE, S.J., 1871. A Missionary Cruise in the South Pacific ... in 1870. Sydney, Joseph Cook.
• WILKES, C., 1845. Narrative of the United States Exploring Expedition During the Years 1838, 1839, 1840, 1841, 1842 (5 Volumes). Philadelphia, Lea and Blanchard.
• WOODFORD, C.M., 1916. “On Some Little-known Polynesian Settlements in the Neighbourhood of the Solomon Islands.” The Geographical Journal, 48:26-54.