- 53

Classic scholarship on the nature of population regulation in pre-industrial settings drew heavily on historic Pacific Island cases to demonstrate that populations can provision a variety of effective solutions to problems of “overpopulation” (e.g., Carr-Saunders 1922; Krzywicki 1934). Raymond Firth's (1936) Tikopian study provided an early and important contextualisation of the ways in which island populations could, and did, actively control population size through such mechanisms as infanticide, induced abortion, celibacy and emigration. As such, the “Pacific Island model” has been extremely important in influencing the way populations in circumscribed environments are conceptualised in demography, and has provided a foundation for the modern theory of pre-industrial fertility control (e.g., Caldwell et al. 1987).

However, much of the current demographic commentary in the Pacific has a very different tone, focusing on the apparent lack of internal checks to population and posed as applying to a very different context — that of a rapidly modernising region. In the last several decades there has been (ever-increasing) expressions of concern about the “problem” of “overpopulation” resulting from high rates of intrinsic growth in many parts of the region (Ahlburg 1987, 1989; Connell 1986). For example, Ahlburg states:

The projected estimates for the region are very high, ranging between 50 and 100 per cent increase in size in the next two decades (Ahlburg 1987).

The failure of family planning regimes to reduce fertility and forestall rapid population growth in parts of the Pacific has often been cited as primary factor in this “overpopulation”. Family planning programme impacts on fertility have been extremely uneven between countries. Fiji, for example, is considered a family planning “success story”, but other nations, especially the smaller and “less developed” atoll nations, are considered to be relative “failures” — where - 54 acceptance rates have been dramatically lower than Governments expected and desired. This is despite biomedical contraceptive methods being available free to users and widely available in many countries (e.g., Lucas and Ware 1981; McMurray and Lucas 1990; Levy et al. 1988).

Anthropological studies in a wide variety of cultural settings have demonstrated an intimate and important relationship between women's contracepting behaviour and their ideas about physiological and reproductive process (e.g., Ibanez-Novion 1980; MacCormack 1985; Maynard-Tucker 1989; Nichter and Nichter 1987; Shedlin and Hollerbach 1981; and offerings in Newman 1985). A recent survey of the state of family planning in the Pacific by McMurray and Lucas (1990) underscores how little this has pervaded family planning research and policy in the region, even though it has the potential to throw light on vital aspects of family planning behaviour which are not discernible through other approaches.

This paper examines women's knowledge of reproductive physiology and anatomy and its relationship to patterns of biomedical contraceptive use in a rural Micronesian population. An aim is to show how I Kiribati women's perceptions of the anatomy and physiology of their own bodies and the expected or perceived health outcomes of using clinic contraceptives can have a pronounced impact on the acceptance, continuance and use-effectiveness of clinic contraceptive methods, and thus on community-level fertility performance and, ultimately, on population growth rates.

This report is based on field work conducted by the author on the equatorial atoll of Butaritari (Fig. 1) in the north of the Tungaru (Gilbert) chain of the Republic of Kiribati between mid-1990 and mid-1991. Butaritari has a predominantly subsistence fishing-horticultural economy and most of the population (79%) is Catholic in religious affiliation. At the end of 1990 the population of the island was approximately 3800, distributed across eight main villages in an area of 13.5 km2. Proportional mortality from infectious disease is reportedly higher in Kiribati than elsewhere in Micronesia-Polynesia (Taylor et al. 1989), and this is reflected in infant mortality rates as high as 10 per cent on some atolls, including Butaritari (see Brewis 1993).

Fertility is currently and historically high in Kiribati and on Butaritari. The annual rate of natural increase in the last intercensual period (1985-1990) was 2.24% nationally (Rouatu 1991). Opportunities for permanent emigration are limited, but there has been notable rural-urban drift to the central administrative capital of Tarawa over the last several decades. In 1990, Butaritari had a de facto population of 280 persons per km2, and Government projections from this figure estimate that, if current vital rates prevail, density will rise to over 600 persons

- 55

Figure 1: Location Map of Butaritari and Republic of Kiribati.

- 56

per km2 on the atoll by the year 2025 (Tiroa 1990).

According to an island census conducted at the outset of the field season, ever-married women over the age of 15 living on Butaritari in late 1990 (N=791) have a mean of 4.65 live births. Those who have reached menopause (N=231) have a mean completed family size of 7.13. There is an exceptionally low primary infertility rate in this population — mainly due to a relative historical absence of sexually transmitted diseases (Brewis 1992a, 1993); only 1.29% of ever-married postmenopausal women alive at the end of 1990 have not produced at least one live birth. Further, virtually all men and women marry at least once; at the end of 1990, 96 per cent of women living on the atoll 25 years of age or older reported being currently or previously married.

In 1972, only four years after an active and positive national family planning campaign was instituted, reportedly over one-third of all I Kiribati women aged 15-44 years had adopted and continued to use a biomedical contraceptive method (intra-uterine device, oral or injectable). It was then thought “…a fairly high rate of family planning continuance in all age groups is, unless something goes fairly wrong, likely….The colony may therefore have set a precedent” (Pitchford 1972:12).

It seems, in retrospect, something did go “fairly wrong”. This “failure” is especially notable, given that Government planning policies since that time have made continued efforts to lower fertility through the increase of family planning coverage (e.g., Government of Kiribati 1978, 1982, 1986). By 1979, it was reported that less than 16 per cent of I Kiribati women aged 15-49 years were using any contraceptive method, including permanent female sterilisation by tubal ligation (Government of Kiribati 1989). Although a sustained effort in family planning method deliverance and public education has raised national family planning acceptance and continuance rates in the last decade to about a quarter of women aged 15-49 years nationally (Government of Kiribati 1989), it has been noted there has also been a trend over the last 20 years towards a proportional increase in the use of oral contraceptives and Depo-Provera injectables over the use of intra-uterine devices (IUDs), the former methods being less efficacious if used inconsistently (Lucas and Ware 1981:311).

Biomedical contraceptives are currently widely available on Butaritari atoll, distributed by local nurse-aides and nurses at the clinics situated in the main villages. Methods consistently available at each clinic include Eugynon oral contraceptives, Depo-Provera injectables (depot medroxyprogesterone acetate), and male condoms. Depo-Provera is the method most actively promoted by clinic staff, who find it easiest to dispense, explain to users and follow up. Condoms are available in the clinic but are not currently being dispensed and there are no current or previous users in the study sample (see below). At the time of field work, IUD insertion was not being done as no physician was resident on - 57 the island. All clinic-based methods are provided free to users, and clinics are centrally located and easily accessible to everyone on the island. Catholic Church personnel have intermittently conducted Billings' method “natural” family planning classes (based on mucal symptomology) over the last two decades, although none of these classes took place during the field work season, to my knowledge.

In addition to this range of biomedical options, Butaritari women have a number of “traditional” family planning options available if they elect to seek and employ ethnomedical alternatives. “Traditional” methods of birth control include a form of coitus interruptus/withdrawal, use of herbal medicational drinks, and abdominal or cervical manipulation or retroversion of reproductive organs to prevent conception. These latter methods are discussed in greater detail below, but all these are very rarely practised at present and are stated to have never been practised by any of the premenopausal women in the study sample. 1 According to Grimble (1921), earlier this century I Kiribati subscribed to a prolonged period of sexual abstinence between a birth and the time a child is weaned. As is reported elsewhere (Brewis 1993; Brewis and Regmi 1993), this post-partum ideology still prevails, but many couples have intercourse very soon after a birth in order to negotiate the perceived health risk to a child of resuming coitus before weaning. 2 Induced abortion by “traditional” means still occurs, although its use is limited to very rare cases of unmarried women attempting to keep a pregnancy secret. As such, it cannot be considered a ready family planning option. 3

Despite this range of biomedical and traditional family planning options, fertility levels on Butaritari remain high. Certainly, biomedically based family planning programmes are not having a substantial demographic impact — at least not in any manner consistent with national Government aspirations for fertility reduction. In fact, Butaritari fertility is barely lower now than in the decades preceding the introduction of clinic contraceptive methods (see Brewis 1992b), and current Butaritari fertility measures rank high even when compared with other groups thought to practise no clinic contraception at all (see Campbell and Wood 1988: Appendix A; Bentley et al. 1992) — although they are on par with contemporary rural Marshallese populations (Levy et al. 1988).

Reproductive histories were taken by the author and a local assistant in a single village, the largest of eight villages on the atoll. In total, 203 women presented reproductive histories, which constituted 87.1% of all women aged over 15 currently resident in the village. 4 Reproductive history records were supplemented by, and cross-checked with, vital events recorded in Government birth, death and marriage registers, and medical and vital events recorded in - 58 village clinic records. These verified reproductive histories provide the demographic data presented in this paper. The sample proved to be adequately representative of the total female population of the island with respect to age distribution, marital status, religious affiliation and general fertility measures (e.g., completed family size, mean number of live births, woman-child ratios) (see Brewis 1992b). Of the final sample of 201 women, 169 (84.08%) are Catholic, a further 10.95 per cent are members of the Kiribati Protestant Church and the remainder are Seventh Day Adventist (n=5), Baha'i (n=4) and Mormon (n=1). Ever-married women constitute 85.6 per cent of the sample, 72.6 per cent of the sample are currently married.

Thirty-four of the women who provided reproductive histories gave subsequent semistructured interviews in the areas of reproductive ethnophysiology and ethnoanatomy, family planning beliefs and behaviours, and women' s health symptomology and treatment. This subsample of women was non-randomly selected, participants being chosen so to provide a broad cross-section of different ages and marital-reproductive histories, while selecting for women with whom I developed a strong personal rapport or who had been most responsive in the initial reproductive history interviews. In specific discussions of reproductive physiology, women were asked to draw internal and external anatomy and explain physiological processes using the blank outline of a mature female figure, following the method detailed by Shedlin (1979) and MacCormack (1988). In addition, eight ethnomedical healers and/or midwives and three female nurse-aides were interviewed on a broad range of fertility-associated topics. Wherever possible, I took part in, or observed, women's healing and fertility-associated events, including those at the village clinic and ethnomedical treatments by specialists in their own or patients' homes. The discussion of notions of reproductive physiology and health is based on these data.

The reproductive status of currently married, premenopausal women in the sample (N=116) is presented in Table 1. Of these women, about 70 per cent (n=82) can be considered currently at risk of becoming pregnant — that is, are sexually active, reportedly not pregnant and had a menstrual event in the six weeks before being interviewed. Of those who could be considered at risk of pregnancy, 18.29 per cent (n=15) are currently using a clinic-based method of pregnancy avoidance (seven using orals, eight using Depo-Provera). Biomedical contraceptive method acceptance rates at the time of interview can therefore be considered low for women in the sample, and assumably for the island generally. An additional 8.54 per cent (n=7) of these potential users are employing a nonclinic method, in all cases Billings' rhythm method.

When compared with the number of this same group of women who state

- 59

cont — currently using a clinic method of contraception

preg — currently pregnant by self-report

amen — not pregnant, no menstrual event in the six weeks before interview, not lactating

lacyc — not pregnant, menstrual event occured in the six weeks before interview, currently lactating

laamen — not pregnant, no menstrual event in the six weeks before interview, currently lactating

cycl — not pregnant, menstrual event occured in the six weeks before interview, not lactating

%susc — percentage of women in age class who are susceptible, i.e., who are not pregnant and report a menstrual event in the six weeks before interview

mean births — mean number of live births

- 60

they have no specific plans to ever have any more children, as set out in Table 2, there is a marked difference between the level of contraceptive practice and immediate desire to limit family size — the so-called “unmet need” (Bongaarts 1991). That is, 35.34 per cent of currently married, premenopausal women state they wish never to have any more children and a further 18.97 per cent have no immediate desire for more children. The remaining 45.69 per cent includes all those who consider they may at any future time desire another child, and thus represents those who currently wish to space children as well as those actively attempting pregnancy. Despite this being less than half of all the women, only just over a quarter of all eligible women are currently contracepting by any method.

This gap between women's immediate family size desires and women's failure to contracept has been noted by other researchers working in Kiribati, and has been explained variously by “male factors” and “religious resistance” (e.g., McMurray and Lucas 1990:24; Pitchford 1977). From the Butaritari data, neither of these factors specifically explains the low method use-rates in relation to expressed desires to limit family size. In respect to “male factors”, there is a high degree of concordance between women's statements of desire for another child and the views they believe are held by their partners (see Table 2). That is, 93.1% of women state that their partner does not disagree with their immediate family - 61 size goals. With regard to “religious resistance”, there is no statistically significant difference in current fertility between Catholics and non-Catholics in this sample. 5

The methods which ever-married women claim to have ever used for either birth spacing or birth stopping are set out in Table 3. Of women who are currently premenopausal and who have had at least one child survive to the age of 12 months — and could, therefore, presently be considered a potential method user — just under half (44.97%) have never employed any overt method of contraception. 6 Of those accepting methods at any time previously, 43 per cent have used Depo-Provera, 23.6 per cent have used orals and 7.5 per cent have used an IUD. Billings' method, however, has had the greatest proportion of acceptors, with 47.3 per cent of ever-contraceptors using the method at some time. Of the women who have ever employed any form of contraceptive, almost 60 per cent have at some time elected to use a clinic-based method.

- 62

Given this situation, outright method resistance does not explain the current low levels of use of biomedical contraceptive options by these women. Rather, contraceptive behaviour on Butaritari appears to be characterised by high levels of method discontinuance and intermittent or ineffective use. This observation has been made informally by Lloyd and Winn (1985:39), who have argued national user-rate figures are somewhat deceptive given the likelihood of high rates of ineffective method use.

Certainly, the factors underlying contraceptive decision-making in Butaritari women are exceedingly complex and relate to a vast array of factors which cannot be described in the space here (but see Brewis 1991, 1992b). However, the reasons that women cite for discontinuance of various methods, as set out in Table 4, focus substantively on perceived method side-effects. Further, it emerged in the course of interviewing that not only are women very concerned with their own experiences of method side-effects, but also that these concerns are passed on to non-users. That is, rumours of unpleasant method side-effects appear to be inhibiting method acceptance in women who might otherwise elect to limit family size through biomedical options. It is in this area that investigation of cultural constructions of physiological process and anatomy can prove enlightening to explaining how uptake “failures” and discontinuance patterns may be reflection of women's interpretations of the action of various methods on their own bodies. For example, the idea that IUDs can dislodge and travel through a woman's body and lodge elsewhere with dangerous health implications has been expressed in a wide variety of cultural settings and linked to seemingly insurmountable resistance to the acceptance of IUDs as viable contraceptive options (e.g., Maynard-Tucker 1986; Shedlin 1977). The following section describes women's ethnophysiology and ethnoanatomy as it is tied to perceptions of the use of biomedical contraceptive options.

The I Kiribati women's model of the physiology of conception, pregnancy, and reproductive health is not particularly complex. What follows is an outline based on consistent themes which emerged during interviews.

Both lay and local specialist (traditional healer) understanding of internal female reproductive anatomy focuses on two primary organs — te ririniman (the cervix) and te kai (roughly analogous to the uterus). Te kai is the central reproductive organ, which normally sits directly below, and under, the navel. It is a spherical or ovoid organ of about fist-size. The kai is both the source or store-place of menstrual blood and the site of foetal formation and development. It is an unanchored organ, and it can move in any direction within the abdomen from

- 63

- 64

its correct and healthy central position under the navel. This movement is most commonly caused by sudden physical bodily strain (such as by heavy lifting or falls) or abdominal heating (kabuebue). The misalignment of te kai can cause a variety of disorders — depending on where it comes to rest. The most common symptoms of movement are failure to conceive, amenorrhoea or menstrual irregularity and localised pain. The common treatment for a “crooked” or displaced kai is abdominal pressuring and massage, usually in combination with external “cooling” therapy. All these procedures are conducted by a specialist.

Te ririniman is described as a hard, round organ at the top of the vagina which acts as a doorway to the inner body of a woman. It is cited as the source of vaginal secretions. It is located away from te kai — the two do not touch and are separated by a “liquid” expanse.

In the I Kiribati women's model, intercourse is necessary for “conception” to occur and only one sexual act is sufficient for pregnancy. Although most women state that pregnancy can occur at any point of the menstrual cycle, about one-third believe that there is a specific fecund time — most often said to be the few days immediately following menstruation. According to the ethnoanatomical model, the penis normally sits directly at the cervical opening during sex. This allows the entry of the male ejaculate (tari) into the inner body of the woman and also promotes female orgasm as the penis knocks against the ririniman. For “conception” to occur, both partners should orgasm; to be most effective, this should occur simultaneously. The rhythmic contractions of female orgasm are said to be the spasmodic opening and closing of the ririniman. This is why it is considered difficult to become pregnant in the absence of female orgasm: it is the opening of the ririniman which allows the male tari to enter the woman's body and thus for “conception” to occur. 7 After the seminal fluid enters the woman's body, it mixes with her procreative fluid or “juice”, also referred to as her tari. This composite mix of complementary male and female elements then migrates directly upward and enters te kai. 8 Once settled in te kai, this mixture begins to feed on retained menstrual blood and continues to use this blood as the food source for the duration of the pregnancy. Without this menstrual blood retained in pregnancy the foetus cannot develop.

Several specialists in the village have practised on occasion some form of uterine or cervical manipulation to either encourage or discourage pregnancy. The practice is used very rarely for contraceptive purposes, and much more commonly for infertility disorders. For birth control purposes, abdominal massage is practised to move te kai off-centre (as, if te kai is out of place, no pregnancy will ensue because this mixture will not travel into te kai and will have “no place” to form a foetus). Alternatively, the cervix is manipulated so it sits “above” the head of the penis during intercourse so the man's tari will not be able to enter the woman's body.

- 65

Women's identification of their own reproductive health focuses on three primary symptoms. These are regular menstruation, low level of vaginal secretions, and sex organs which do not smell distasteful or unclean. Of these, menstrual regularity is the most worrying in terms of perceived health risk, and the menstruation event is seen as health-preserving. Menstrual regularity and the menstrual event are distinguished as more important to an attractive and healthy female sexuality, although they also figure symptomatically in the identification of reproductive ill-health. To the list of reproductive ill-health symptomology must be added the misplaced kai and kabuebue (fever or heat), the latter being the primary and single most important disease symptom in Kiribati ethnomedicine. 9

One of the most feared reproductive conditions women may experience is te tibu, characterised by the prolonged “retention” of menstrual blood. Women repeatedly note that, when te kai is out of line, the menstrual blood may be held in the organ and the woman is at risk for this life-threatening condition. The retained blood is said to eventually swell the “stomach” and kill the woman. If a woman has not menstruated, or has only menstruated lightly for some months and suspects stomach swelling (and has no other pregnancy symptoms), she is highly motivated to seek specialist care. The healer uses gentle downwards massage to “release” the blood and to replace the kai in its proper position. In this way, certain forms of amenorrhoea are considered symptomatic of a potentially fatal condition.

Ten rara is another feared women's reproductive condition, represented symptomatically as heavy, prolonged bleeding from the vagina (whether menstrual or other). This is said to happen when the cervix is “not properly closed”. It is treated through herbal drinks or through massage but, if not adequately handled, it can lead to death when the woman is “left with no blood”. Heavy prolonged menses is sometimes interpreted as a ten rara illness, where the cervix has failed to close properly and blood is continuing to flow out of the body in an uncontrolled manner.

Other reproductive conditions which are of concern to Butaritari women are those in which watery, foul-smelling vaginal discharge occurs. Such conditions are considered symptomatic of either water entering the reproductive organs (such as is said to enter through an open cervix if a woman bathes in the sea during menstruation), or of her own tari or sex organs turning rotten (mka). Rotting organs is thought to be one result of the rare practice of vaginal and cervical cauterisation, or kotokoto. 10

Possibly the most dreaded reproductive condition of which Butaritari women are aware is te kiriti (lit. grease). This condition is associated with pregnancy and is said to mimic pregnancy symptomology—including increasing - 66 stomach girth, tender breasts, pregnancy craving and amenorrhoea, but which often results in the death of the woman. 11

Generalised vaginal itching or discomfort is another symptom of reproductive ill-health, and it is considered most commonly the result of an overly wet or “unclean” vagina. It is both prevented and cured by manually removing all mucus from the vagina once or twice each day and after sex; a practice habitually undertaken by married women.

In discussing reproductive ill-health, it becomes apparent that many of the conditions that may be signalled by reproductive symptomology (such as amenorrhoea or vaginal discharge) are perceived as possibly life-threatening. In fact, every woman could name several individuals they considered to have died of these conditions over the last several years. It is from these conditions, as well as birth and pregnancy complications, that most adult women are thought to die. And, alternatively, it is by avoiding these conditions, women believe, that a woman gets to live a long and healthy life.

It is within this framework of concern over reproductive symptomology that the following discussion of perceptions of clinic contraceptive method use is gauged — that is, the way in which the methods “work” in the body, their perceived side-effects and the relationship with different conditions of reproductive ill-health.

Women are generally unsure how orals affect fertility, although some note that the pill may kill the female tari by heating the abdominal region. As such, it may have a dangerous heating effect on the body and could therefore be a precipitating factor in an enormous array of diseases. For this reason, orals are particularly avoided by lactating women — who are concerned that the heating effect will taint breast-milk and make the breast-feeding child ill with diarrhoea and vomiting. Women are satisfied, however, with the menstrual regulation which accompanies pill use, which is an indicator of reproductive health.

One mentioned side-effect of the pill is an increase in feelings of vaginal wetness and in mucus production and increase in unpleasant vaginal odour, which are considered by women to be sexually undesirable to their partners. Other occasionally mentioned side-effects include fatigue and dizziness, although these were not considered of great concern by those reporting them.

When the Lippes loop IUD was first introduced in the Butaritari clinic, women were extremely enthusiastic about the new method. At present, however, the “loop” is considered a dangerous and unpleasant contraceptive option. One - 67 aspect of the fear associated with the loop is the story, which has now circulated for several years, of a woman who died after a loop “travelled into her stomach”. This story has now become firmly entrenched in local knowledge, and most women interviewed can recount a version. This particular story has been crucial in fuelling fear about using this method, encouraging discontinuance and a decline in subsequent acceptors. Women say that the loop sits in, or on top of, the cervix. It works by killing either the man's or the woman's tari. As the area above where the loop sits is empty of all organs and filled with liquid, there is the possibility that the loop can become unattached, travel freely upward, and damage internal organs. If a woman becomes pregnant while using a loop, it is thought to be potentially damaging to the foetus — causing congenital deformations such as harelip and cleft palate or birthmarks. An “escaping” loop has also been linked on occasion to the occurrence of ten rara, although informants were unable to describe the exact nature of the causal connection.

Another problem described with the loop is the irregularity of menstruation that many women find unacceptable as a side-effect. As described above, menstrual regularity is a very important aspect of feminine health, and irregular bleeding is seen as a possible symptom of te tibu. Therefore, any woman noting these side-effects is highly motivated to discontinue method use.

Several women had loops spontaneously release, and this was very frightening for them as they were concerned about bleeding to death and felt this was also evidence of the mobility of the device in the body. In other cases, women with te kiriti had connected this condition with previous use of the loop. Thus, there is a perceived connection between this condition and loop use, which had made some wary of the methods given how fearful Butaritari women are of te kiriti.

Depo-Provera is the most popular clinic contraceptive method used on Butaritari at present, in part because it is the method most aggressively promoted by clinic staff. The problem most often cited with this method is the possibility of te tibu occurring, as menstrual blood is thought not to be released regularly. Women state that the possibility of te tibu is much higher in te iti than the loop because weight increases are noted by women using this method. It is the weight gain, rather than the cessation or irregularity of menstruation, that is the greatest concern to users. This is because the weight gain is considered a clear indicator that te tibu is in progress, as the kai becomes full and heavy with “retained” menstrual blood. Because te tibu is thought of as a potentially fatal condition, women who perceive body changes indicative of te tibu are strongly motivated to discontinue this method.

- 68

It is well established in the demographic-anthropological literature that lay notions of reproductive physiology and anatomy can have substantial influence on the ways in which community members perceive, accept and use biomedical contraceptive options. As such, this relationship has the potential to significantly affect the demographic impact of biomedically based family planning programmes on community fertility. When clinic contraception was first made available to Butaritari women 25 years ago, acceptance rates were significantly higher than they are today. The decline in the interim is not associated with either increasing “religious resistance” or marked increases in desired family size. In the Butaritari case, women are interpreting contraceptive methods within an ethnomedical-ethnoanatomical framework, and this is having a substantial impact on the ways they use or, rather, do not use biomedical contraceptive options. This is one case in which increasing knowledge of and experience with biomedical contraceptives has decreased, rather than increased, acceptance. Certainly, the widespread availability of effective, free biomedical contraceptive options to sexually active women with birth-spacing or birth-stopping desires proves insufficient to guarantee any substantive impact on community fertility in this Butaritari case. In summary, given local definitions of an “unmet need” for contraception on Butaritari, dissatisfaction with biomedical method options is crucial.

The data on ethnophysiological notions and method use and discontinuance provide but one small key to explaining the complex interplay of factors which affect the demographic impact of family planning programmes on Butaritari. This is, however, an area which has been long overlooked in demographic analyses of family planning behaviour and impact in the Pacific region, and one about which further culturally-specific examples could prove illuminating. In surveys of the literature, I found only one study from the region in which the complex relationship between ethnophysiology and contraceptive behaviour was tackled direct — this is in the work of Anne Chambers (1986) based on research in Tuvalu. As with this present study, Chambers found that resistance to clinic contraceptive methods is closely tied to notions of reproductive ethnophysiology among Nanumean women. 12

The investigation of ethnoanatomical models is but one way to chart the extremely complex relationship between the cultural construction of reproductive events and biomedical contraceptive behaviour and its demographic outcome for Pacific Island populations. However, given the priority that Butaritari women afford contraceptive method side-effects in their contracepting decisions, and the divergence of their own models from biomedical models of both physiology and side-effect symptomology, such investigations may prove fruitful in helping to place the needs and desires of potential users in greater concordance with those of their Governments.

- 69

A Mellon Foundation Fellowship held in the Population Studies and Training Center provided the opportunity to prepare this material. I thank Monica Das Gupta, Judith Huntsman, Bernd Lambert, Lucile Newman, Mark Nichter and Jane H. Underwood for their careful attention and comments, and the women of Butaritari for their friendship, trust and humour. Research in Kiribati was supported financially by the New Zealand Health Research Council and through a Rockefeller Foundation Predoctoral Fellowship in the Population Sciences.

- 70 Page of endnotes

- 71

• Ahlburg, D., 1989. Demographic Pressures on Health, Education and Employment Resources in the South Pacific Region. Pacific Studies, 12:23-31.
• —— 1987. Population Forecasts for the South Pacific Nations using Auto-regressive Models: 1985-2000. Journal of the Australian Population Association, 4:157-67.

• Bagshawe, K., 1969. Choriocarcinoma. Baltimore: Williams and Wilkins.
• Bentley, G., T. Goldberg and G. Jasienska, 1992. The Fertility of Agricultural and Non-agricultural Traditional Societies. Population Issues Research Center, Pennsylvania State University, Working Paper, No. 1992-02.
• Best, E., 1924. The Maori. Wellington: Board of Maori Ethnological Research.
• Bongaarts, J., 1991. The KAP-Gap and the Unmet Need for Contraception. Population and Development Review, 17:293-313.
• Brewis, A., 1991. Family Planning Behaviour in Cultural and Demographic Context: A Report to the Kiribati Ministry of Health and Family Planning. Monograph, Tarawa, Kiribati.
• —— 1992a. Sexually-transmitted Disease Risk in a Micronesian Atoll Population. Health Transition Review, 2:195-213.
• —— 1992b. Age and Infertility: An Ethnodemographic Investigation from Butaritari Atoll, Kiribati. Unpublished Ph.D. thesis, University of Arizona.
• —— 1993. Age and Infertility in a Micronesian Atoll Population. Human Biology, 65:593-609.
• —— and G. Regmi, 1993. Post Partum Amenorrhoea Differentials and Patterning in a Rural Pacific Island Population. Population Studies and Training Center Working Paper. Series 93-05. Brown University.

• Caldwell, J., P. Caldwell and B. Caldwell, 1987. Anthropology and Demography: The Mutual Reinforcement of Speculation and Research. Current Anthropology, 28:25-43.
• Campbell, K. and J. Wood, 1988. Fertility in Traditional Societies., in P. Diggory, M. Potts and S. Teper (eds), Natural Human Fertility: Social and Biological Dimensions. London: Macmillan, pp.39-69.
• Carr-Saunders, A.M., 1922. The Population Problem: A Study in Human Evolution. Oxford: Clarendon.
• Caughey, F. B., 1971. The Social Context of Pregnancy and Birth on Uman, Truk. Unpublished M.Sc. thesis, University of Pennsylvania.

- 72

• Chambers, A., 1986. Reproduction in Nanumea (Tuvalu): An Ethnography of Fertility and Birth. Working Papers in Anthropology, 72. Anthropology Department, University of Auckland.
• Connell, J., 1986. Population, Migration and Problems of Atoll Development in the South Pacific. Pacific Studies, 9:41-58.

• Firth, R., 1957. We, the Tikopia. 2nd edit. London: Allen and Unwin.
• Fischer, A., 1963. Reproduction on Truk. Ethnology, 2:526-50.
• Fitzgerald, M., 1989. Modernisation and the Menstrual Experience among Samoans. Unpublished Ph.D. thesis, University of Hawai'i.
• —— 1990. The Interplay of Culture and Symptoms: Menstrual Symptoms among Samoans. Medical Anthropology, 12:145-67.
• Franklin, R., 1965. Leprosy in the Gilbert and Ellice Islands. Unpublished Dip. T.P.H. thesis, London School of Hygiene and Public Health.

• Government of Kiribati, 1978. Kiribati Fourth National Development Plan, 1978-1982. Bairiki, Tarawa: Ministry of Finance.
• —— 1982. Kiribati Fifth National Development Plan, 1983-1986. Bairiki, Tarawa: Ministry of Finance.
• —— 1986. Kiribati Sixth National Development Plan, 1987-1991. Bairiki, Tarawa: Ministry of Finance.
• —— 1989. Kiribati 1979-1987: A Compendium of Statistics. Bairiki, Tarawa: Statistics Office.
• Grimble, A., 1921. From Birth to Death in the Gilbert Islands. Journal of the Royal Anthropological Institute, 51:21-54.
• ——1930. Unpublished report of population, August 30.

• Ibanez-Novion, M.A., 1980. The Pill and Traditional Knowledge of the Human Body: A Brazilian Case Study, in G. I. Zatuchni, M. Labbok, and J. J. Sciarra (eds), Research Frontiers in Fertility Regulations. New York: Harper and Row, pp.4-72.

• Jacobs, P., P. Hunt, J. Matsuura, C. Wilson, and A. Szulman, 1982. Complete and Partial Hydatidiform Mole in Hawaii: Cytogenetics, Morphology and Epidemiology. British Journal of Obstetrics and Gynaecology, 89:258-66.

• Kinloch, P., 1984. Midwives and Midwifery in Samoa, in C. Parsons (ed.), Healing Practices in the South Pacific. Honolulu: IPS, pp. 199-212.
• Krzywicki, L., 1934. Primitive Society and its Vital Statistics. London: Macmillan.

• Lambert, S., 1924. Health Survey of the Gilbert and Ellice Islands: With Special Reference to Hookworm Infestation. Suva: Western Pacific High Commission.
• Levy, S., R. Taylor, I. Higgins, and D. Grafton-Wasserman, 1988. Fertility and Contraception in the Marshall Islands. Studies in Family Planning, 19:179-85.
• Lloyd C., and M. Winn, 1985. Final Report on Women's Reproductive Health and Family Planning in the South Pacific. Sydney: Australian Federation of Family Planning Associations.
• Lucas D., and H. Ware, 1981. Fertility and Family Planning in the South Pacific. Studies in Family Planning, 12:303-15.

• MacCormack, C., 1988. Lay Concepts of Reproductive Physiology Related to

- 73

• Contraceptive Use: A Method of Investigation, in J. C. Caldwell, A.G. Hill and V. J. Hull (eds), Micro-Approaches to Demographic Research. London and New York: Kegan Paul, pp. 441-8.
• —— 1985. Lay Concepts Affecting Utilisation of Family Planning Services in Jamaica. Journal of Tropical Medicine and Hygiene, 88:281-5.

• Macdonald, J., 1991. Women of Tikopia. Unpublished Ph.D thesis, University of Auckland.
• McMurray, C., and D. Lucas, 1990. Fertility and Family Planning in the South Pacific. Islands/Australia Working Paper No. 90/10. National Centre for Development Studies: Australian National University.
• Macpherson, C., and L. Macpherson, 1990. Samoan Medical Belief and Practice. Auckland: University of Auckland Press.
• Maynard-Tucker, G., 1986. Barriers to Modern Contraceptive Use in Rural Peru. Studies in Family Planning, 17:308-316.
• —— 1989. Knowledge of Reproductive Physiology and Modern Contraceptives in Rural Peru. Studies in Family Planning, 20:215-24.

• Newman, L., 1985. Women's Medicine: A Cross-Cultural Study of Indigenous Fertility Regulation. New Brunswick: Rutgers University Press.
• Nichter, M., and M. Nichter, 1987. Cultural Notions of Fertility in South Asia and their Impact on Sri Lankan Family Planning Practices. Human Organisation, 46:18-28.

• Oliver, D., 1974. Ancient Tahitian Society. Honolulu: University of Hawaii Press.

• Pitchford, J., 1977. Family Planning in the Gilbert and Ellice Islands. South Pacific Bulletin, 22:13-16.
• —— 1972. Factors Affecting Fertility on the Outer Gilbert and Ellice Islands: A Pilot Study. Tarawa: Family Planning Association.
• Pukui, M. K., 1942. Hawaiian Beliefs and Customs during Birth, Infancy and Childhood. Bishop Museum Occasional Papers 17. Honolulu: Bishop Museum.

• Rouatu, I., 1991. Kiribati Population Census — 1990 Summary Results. Unpublished manuscript, Kiribati Department of Statistics.

• Shedlin, M., 1977. Body Image and Contraceptive Acceptability in a Mexican Community. Paper presented at the Annual Meeting of the American Association for Anthropology, Houston, TX, November 29-December 3, 1977.
• Shedlin, M. Goldzieher, 1979. Assessment of Body Concepts and Beliefs Regarding Reproductive Physiology. Studies in Family Planning, 10:393-7.
• —— and P. E. Hollerbach, 1981. Modern and Traditional Fertility Regulation in a Mexican Community: The Process of Decision-making. Studies in Family Planning, 2:278-96.
• Shore, B., 1981. Sexuality and Gender in Samoa: Conceptions and Missed Conceptions, in S. Ortner and H. Whitehead (eds), Sexual Meanings: The Cultural Construction of Gender and Sexuality. Cambridge: Cambridge University Press, pp. 192-215.

• Taylor, R., N. Davis Lewis and S. Levy, 1989. Societies in Transition: Mortality Patterns in Pacific Island Societies. American Journal of Epidemiology, 18:634-46.

- 74

• Thomson, B., 1908. The Fijians. London: Macmillan.
• Tiroa, T., 1990. Estimated Populations of Kiribati for the Years 1990 and Beyond. Unpublished manuscript, Kiribati Department of Statistics.

• Wilkes, C., 1845. Narrative of the U.S. Exploring Expedition During the Years 1832-1842. Philadelphia: Lea and Blanchard.