Fraternal Polyandry and Fertility in a 
  High Himalayan Valley in Northwest Nepal

  Melvyn C. Goldstein (1)

  The manner in which cultural factors affect fertility has become an important
  area of anthropological concern. In this article, it will be shown that while
  fraternal polyandry does not affect individual fertility it does have a significant
  depressing effect on aggregate fertility and functions, unperceived and unin-
  tended, as an important mechanism for reducing population growth.
  KEY WORDS: fertility; polyandry; Nepal; Tibetans.

  	Whereas early proponents of the demographic transition theory ignored
  cultural variables and contended that preindustrial societies existed in a kind of
  "natural" homeostasis in which natural (uncontrolled) fertility was balanced by
  high mortality, this is no longer accepted today. The importance of cultural
  factors in both the theoretical and applied areas of demography has become
  widely acknowledged and, within anthropology, a subdisciplinary focus is emerg-
  ing to deal with this under the rubric of population anthropology. In this ap-
  proach, the relationships between cultural and demographic variables are seen as
  complex, reciprocal, and often very subtle. This article deals with one such
  relationship by analyzing how fraternal polyandry affects fertility among a tradi-
  tional, preindustrial population in northwest Nepal. It also suggests, from a
  broader ecological perspective, how polyandry affects the overall adaptation of 
  the population to its environment.
  	Of all the forms of marriage and family, polyandry has remained among
  the least well understood in the anthropological literature. Polyandry is a form 
  of marriage where two or more males share a bride, and fraternal polyandry is a
  variety of polyandry in which the two or males are siblings. Anthropoli-
  cal attemps at explaining this rather rare form of marriage and family have 
  included factors such as female shortages, types of descent and inheritance, the
  maintenance of land holdings intact, and even psychoanalytic variables. How-
  ever, the very obvious and fundamental question of the effect of polyandry on 
  fertility and overall adaptation has been ignored in this literature. Benedict
  (1972: 76), in a survey of social factors affecting fertility, sums up the scant
  data on polyandry as follows:

  	Polyandry is so rare and the information about it so meager, that we can say very
  	little about its effects on fertility. The Toda and the Jaunsari, the two polyan-
  	drous societies listed by Nag, show no difference in the fertility levels of polyan-
  	drous and nonpolyandrous married women.

  	In my own previous research on polyandry (Goldstein, 1971), the question
  of "surplus" women was raised: if a significant proportion of the males are
  involved in polyandrous marriages, one would expect there to be a surplus of
  unmarried females bearring other neutralizing factors such as higher female 
  mortality. Murdock (1949: 25) had earlier touched on the reverse of this issue in
  his classic study, Social Structure, when he commented that "despite the paucity
  of cases there seems reason to assume that polyandry may sometimes be due to a
  scarcity of women resulting from the practice of female infanticide." However,
  the nature of the data available to me previously precluded any satisfactory
  analysis of this issue, although there definitely was no pattern of female infan-
  ticide in Tibetan cultural areas. In 1974, I had the opportunity to undertake
  research on this and related aspects of polyandry as part of a larger study on
  monasticism carried out in the Tibetan-speaking Limi Valley in the northwest
  corner of Nepal; in this artlce, I will try to address these issues by analyzing the
  consequences of fraternal polyandry for fertility and overall adaptation in this
  	Limi is a high mountain valley, running northeast to southwest and con-
  taining three villages. All of these are over 12,000 ft in elevation; Tsand, the
  highest village and the one in which the most detailed work was done, is at
  12,900 ft. While politically a part of Nepal, the people of Limi are linguistically
  and culturally Tibetan. They have an agropastoral type os subsistence economy
  which includes large-scale herding of sheep and yak in addition to agriculture.
  Trade and craft production are also important in their overall adaptation.(2)
  	The three villages in Limi have a total population of 791. This population
  claims to be endogamous. Empirically less than 1% of the marriages have been
  with outsides (either Tibetans or Tibetan-speakers in the Humla Karnali
  ecozone south of Limi). The age-sex distribution of Tsang is presented in Fig.1.
  While 49.3% of the overall population in Tsand is female and 50.7% male, there 
  are, in the reproductive age bracket (15-44), more females than males (76 to 68).


  Contrary to Murdock's suggestion, polyandry in Limi clearly is not the result of
  any shortage of women. Data on migration corroborate this.
  	Limi is not affected by movements of population. There is virtually no
  outmigration. Cases of immigration are limited to several Tibetan nomad refugee
  males who married Limi girls after the political upheaval in Tibet in 1960-1961.
  	As in most areas of Tibetan culture, the largest kinship unit in Limi is the
  corporate family. Marriages are usually patrilocal and either monogamous or 
  polyandrous (fraternal). In Limi I have recorded fraternal polyandrous marriage 
  with up to five brothers sharing a bridge, although normally the number is only
  two or three. Contrary to practices in Tibet, bigenerational polyandry and
  polygyny (father and son or mother and daughter sharing a spouse) are not
  permitted. Plural polyandry or polygynadry (two or more brothers sharing
  more than one wife) is, however, encountered. Also unlike standard Tibetan
  practice is the fact that bilateral cross-cousin marriage (marriage, for a male, to
  his mother's brother's daugher or father's sister's daughter) is esteemed in Limi,
  whereas in Tibet it is considered incestuous. Such cross-cousin marriage may be
  polyandrous or monogamous.

  	Polyandry is the normal form of marriage when a family has more than
  one son. In Tsang, 20 of the 52 marriages recorded were polyandrous. Of the 32
  monogamous marriages, seven were in instances where a family had only one
  son, 23 were cases where brothers split off from previously consummated poly-
  androus marriages, and two were cases where brothers split up before marriage,
  each of them marrying monogamously.
  	The 20 polyandrous mamages included 54 males (seven under 14 years),
  whereas the monogamous marriages included only 32 males. The 20 polyandrous
  marriages contained 23 females(3) thus yielding an average of 2.35 males per
  female. If we assume that without polyandry each of the 23 females would have
  had one husband, there are, then, 31 "excess" males who are absorbed by
  polyandry, i.e., who will not seek a separate bride. This is 21% of the total males
  in Tsang and 33% of the males from 10 to 54 years of age.
  	Polyandry does not seem to have any significant effect on the fertility of
  individual females, although it is very difficult to control for mediating factors
  with respect to this question. A basic problem in ascertaining fertility levels in
  polyandrous vs. monogamous marriages is to establish whether there has been a
  switch in form (usually polyandrous to monogamous) in the developmental
  cycle of each family in the sample. In Limi there were two main sources of
  switching: (1) death of all but one husband in polyandrous marriage and (2)
  splitting off of all but one male in an initially polyandrous arrangement. A
  similar problem arises in cases where jurally polyandrous marriages are sexually
  monogamous because all but one of the males have not yet reached puberty. In
  some cases, such a situation may continue for even a decade.
  	In the following analysis, I have tried to maximize the accuracy of the data
  by eliminating from the calculations all cases which included any of these ob-
  fuscating elements. Polyandrous marriages in which only one male was of re-
  productive age are assigned to the monogamous gategory. The Tsang data sum-
  marized in Table I show that the average number of living children Is 2.9 for



  monogamously married women and 2.7 for polyandrously manied women.(4)
  These differences are not significant and accord with the previously cited data
  on the Toda and Jaunsari with respect to the absence of effect of polyandry on
  individual fertility. However, polyandry does have very important effects on
  fertility when viewed from a population point of view.
  	Without the intervention of cultural or biological mechanisms to reduce
  the number of females, one would expect that a marriage system such as just
  described would generate a sizable number of "surplus" unmarried women. And
  in Limi there are no intervening mechanisms. There is no infanticide of any kind
  and no polygyny (one man having, by himself, two or more wives).(5) As men-
  tioned above, there were in fact more females than males (76 to 68) in the
  reproductive age bracket. Not surprisingly, we also found a significant surplus of
  unmarried women. In Tsang, 31% (or 21 out of 67) of the females of childbear-
  ing age but over 20 (20-49) were unmarried.(6) This figure would be even higher if
  we were to include 18- and l9-year-old females, but, since they still have a fair
  chance for marriage, they have been omitted.
  	The proportion of unmarried females in each age category is indicated in
  Table II. These proportions are extremely high for a preindustrial traditional
  society. For example, in neighboring, nonpolyandrous India (data presented in
  Table III), we find that for the 20-24 year age bracket two of the three Bengali
  groups have no unmarried females and one has only 2% unmarried females. In
  these communities, there are no women over 25 who were not married at least
  once. In actuality, the proportion of unmarried females over 20 years of age in
  Limi is more like the proportions in European industrial countries. Davis and
  Blake (1956: 218), for example, state that the proportion of women past child-
  bearing age who were never married was 26.3% for Ireland (1946), 20.9% for
  Sweden (1945), 20.1% for Switzerland (1941), 16.8% for England and Wales
  (1931), and 13.3% for Belgium (1930).
  	Unmarried females in Limi either (1) continue to live at home (eight
  cases), (2) establish their own separate households (seven cases), or (3) work as
  servants for others (three cases). There is mild social stigma for females who do
  not get married but there is no ostracism or social isolation. There is, moreover,
  a named social category (morang) for unmarried women having independent
  households, and these units have lower tax and corvee responsibilities than the
  other polyandrous and monogamous family units.
  	Not being married, however is not synonymous with exclusion from the
  reproductive pool. Extramarital relationships, if discreet, are tolerated and there-
  fore a number of the so-called unmarried women do in fact have offspring.
  Actually, half of the women in this category had one or more children. However,
  it is critical to note here that the rate of offspring per "unmarried" female is far
  lower than that of married women. The average number of living children in
  1974 for married women was 3.3 per woman, whereas the average for the
  unmarried women aged 20+ was 0.7 per woman (see Table II). As there is no
  infanticide and no significant difference in diet or style of life, the explanation
  of this difference is assumed here to be frequency of coitus, although data are
  not readily available.
  	Aside from factors such as the need for discreetness and the difficulty of
  privacy, there are several important economic motives which seem to restrain
  frequency of coitus. From the male's point of view, illegitimate children are
  expensive. The genitor is jurally responsible for his children, illegitimate or not,
  and is required to provide a variety of items such as a yak, clothes, a sword, and
  often even some plots of land. There is a great deal of open discussion among
  males (particularly the younger ones) about the risks of having affairs with
  unmarried women and this must certainly act as a restraint on the cautious and
  the poor. From the woman's point of view, children are also difficult to support.
  The payments made by the genitor are insufficient to sustain a child and unmar-
  ried females do not have enough land even to support themselves, let alone their
  children. They make up their deficits partly by weaving cloth and selling it but
  mainly by working for others. There is, therefore, a limit to the amount of
  deficit a woman can overcome, since she can only weave so fast and work so
  many days. My contention is that after the second child, and perhaps even after
  the first, the economic pressures become so great that unless there is some
  unusual source of income available, unmarried females voluntarily either assume
  a celibate role or have intercourse at most on infrequent occasions. One well-to-
  do man In Tsang has an open relationship with his unmarried neighbor. He
  supports her and openly treats her children much like his own. He is able to
  dispense with normal discreetness because his wife many years ago had an
  illegitimate child and thus, quite literally, forfeited her right to object to his


  actions. It is significant to note that in this case the unmarried woman (who is
  40 years of age) has four living children. This is above the average for mairied
  women and the only example of an unmarried woman with more than two
  children. In any case, regardless of the cause of this significantly lower rate of
  children for the unmarried woman, its consequences are obvious for the overall
  population picture.
  	Polyandry effectively restricts 31% of the reproductive-age females over 20
  from full participation in childbearing. If this 31% were married and averaged
  the same 3.3 living offspring per female as married Tsang women do, the differ-
  ence between this and the 0.7 rate for unmarried women would mean an in-
  crease of 47 persons or an increase of 16% of the total village population. Over a
  period of several generations (see Fig. 2), the significance of this for population
  growth is impressive. If half of the 47 "extra" people were females and if they
  averaged 3.3 offspring, then the next generation would have an increase of 78
  persons and the following generation an increase of 129 (based solely on the
  projected offspring of the "unmarried" females). Given the fact that the present
  population of Tsang is only 288, it is easy to see how substantial these increases
  would be.
  	Although, previously unmentioned in the anthropological literature on
  marriage and the family, fraternal polyandry (of the Tibetan type) clearly oper-
  ates to depress population growth by producing a residue of females who are only
  marginally involved in reproduction. However, the adaptive significance of poly-
  andry goes beyond this. My further contention is that fraternal polyandry in Limi
  is part of a negative feedback process which operates to adjust, to a degree at
  least, population size to resources. As an introduction to discussion of the nature
  of this feedback loop, brief comments are necessary on the reasons that the
  villagers themselves have for the practice of polyandry. The reader should be
  forewarned, however, that this is a very comphcated subject which can only be
  dealt with here selectively.
  	Individuals in Limi certainly do not embark on polyandrous marriages to
  reduce individual or aggregate fertility levels. No amount of questioning indi-
  cated that any of the subjects saw any connection between polyandry and
  fertility, and the fact that scholars have also not seen this connection is corroba-
  rating evidence for the elusiveness of this relationship. The inhabitants of Limi
  also do not marry polyandrously because they enjoy sharing a wife with other
  siblings or because of any deeply rooted motivational value such as "sibling
  solidarity." As we shall see, a variety of interpersonal tensions and conflicts are
  not uncommon in polyandrous alliances.
  	The subjects' own explanation of their preference for polyandry is highly
  materialistic. They choose fraternal polyandry to preserve the productive re-
  sources of their family units (primarily land and secondarily animals) across
  generations. Polyandry is perceived and consciously selected as a means of
  precluding the division of a family's resources among its male heirs. Since land is
  scarce and most families have less than 1 acre of arable land, the people of Limi
  consider the maintenance of this land intact, i.e., without being split into smaller
  and smaller parcels, a critical factor in sustaining a satisfactory standard of
  living. Fraternal polyandry is the mechanism they use to accomplish this.
  	Polyandry achieves this goal by providing an intrafamilial milieu in which
  discord is minimized by the presence of only one wife on each generational level
  and thus only one set of heirs. It avoids the development of a situation in which
  nuclear family units within the famly consisting of a brother, his wife, and
  children compete with each other. The fact that there is only one wife and one
  set of children (heirs) is believed to be a major deterrent to serious conflict and
  fusion. Whether this belief is correct is a question beyond the scope of this
  article. Suffice it to reiterate here that this is clearly the primary motivating
  factor for polyandry in Limi (and Tibet).(7)
  	But polyandry is not without problems. Because authority (among
  brothers) is customarily exercised by the eldest brother, younger male siblings
  have to subordinate themselves with little hope of ever changing their status.
  When these younger brothers are aggressive and individualistic, intrasibling ten-
  sions and difficulties often occur. Another, very common source of tension in
  polyandrous families is asymmetry in relations between the wife and her
  husbands. While the cultural ideal in Limi is for symmetrical treatment in terms
  of affection and sexual access, deviations from this ideal occur and generate
  intrafamilial tensions, if not outright conflict. Such "deviations" are particularly
  common when there is a sizable age difference between the partners in the
  marriage. It is not uncommon for gaps of 10 years to exist between brothers
  and, since marriages are initiated when the eldest brother is ready for marriage,
  there are also often sizable gaps in age between the wife and the youngest
  brother. Thus a brother who is 10 when a polyandious marriage takes place
  would find that when he is 20 his wife is almost 30, a situation which might be
  perceived by him as unsatisfactory and which might therefore result in a poor
  relationship with the wife in terms of either companionship or sex. Obviously,
  there can be all kinds of variations on this and it suffices here simply to em-
  phasize that while polyandry provides an answer to one type of culturally per-
  ceived problem (albeit one which the subjects see as critical) it does generate
  other types of problems.
  	In a sense, therefore, polyandry is seen as the lesser of two evils. Degrees
  of individual freedom are traded off for degrees of economic security. Polyandry
  flourishes when younger brothers feel they cannot attain a satisfactory standard of
  living if they split off from their family and go it alone. When resources and
  economic opportunities are abundant, however, it is common for younger
  brothers to split off and establish new (neolocal and monogamous) family units.
  The feedback nature of this situation is obvious. An abundance of economic
  opportunities leads to less complete(8) polyandry and more monogamous family
  units. These in turn produce an increase in population by bringing otherwise
  unmarried females fully into the reproductive cycle. Finally, unless there is con-
  tinual expansion of the economic sphere, the increasing population exerts
  greater and greater pressure on resources and this leads to a greater adherence to
  fraternal polyandry. Unperceived by the subjects, then, polyandry functions as a
  sensitive cultural mechanism for adjusting population levels to changes in re-
  source availability and economic productivity.
  	In summary, this article has tried to expand our growing awareness of the
  manner in which cultural institutions act to keep fertility levels lower than the
  biological potential of populations. It has done this by showing briefly the effects
  of polyandry on fertility in one Himalayan population. The article has illustrated
  how polyandry, while not affecting individual fertility of those who are married,
  does significantly affect fertility rates by generating a surplus of unmarried
  females. The article also shows how this consequence is unintended and unper-
  ceived by the subjects, who opt for fraternal polyandry rather than monogamy
  for economic reasons. Finally, it is hypothesized that from a diachronic per-
  spective polyandry in Limi is part of a negative feedback system which func-
  tions to adjust population to changes in resource availability.

  (1)Department of Anthropology, Case Western Reserve University, Cleveland Ohio.
  (2)For a general overview of the adaptation of the population of Limi, see Goldstein (1974,
  (3)The discrepancy between polyandrous marriages and 23 females is explained by the pres-
  	ence of three marriages which consist of two or more brothers and two wives. This is
  	polygynandry, as previously mentioned, and is here included under the larger category
  	"fraternal polyandry."
  (4)These averages are somewhat lower than the total average in Tsang, 3.3 because a number
  	of dubious cases (with respect to polyandry vs. monogamy) have been excluded.
  (5)But there is no cultural prohibition with respect to polygyny.
  (6)There is no simple pattern concerning which females in each generation get married and
  	which do not, and it is clearly not the case that female children of "unmarried" females
  	have a more difficult time finding husbands. In fact, if there is any "class" factor operative,
  	it is associated not with the poor but with the heredity high-status families. Upper-strata
  	marriages require a considerable dowry, whereas poor families without high ascribed status require
  	virtually nothing. Consequently, it is often financiallly difficult for poor but high-status
  	families to marry off daughters, particulary if there is more than one. More important
  	that this factor is the reputation of a female for diligence and hard work, and her personal-
  	ity. Social status is at best a secondary factor.
  (7)Maintenance of an optimum adult labor force is a secondary, although signigicant, motivat-
  	ing factor with respect to polyandry. For a more elaborate examination of polyandry in
  	traditional Tibetan society, see Goldstein (1971).
  (8)The splitting off of one brother does not necessarily end in monogamy when there are
  	three or more brothers, since usually the others remain in the polyandrous alliance.


  	I wish to thank Case Western Reserve University for supporting this
  research by granting me an early sabbatical. I also wish to thank Dean P. R.
  Sharma and Professor A. W. MacDonald of the Nepal and Asian Studies Institute
  of Tribhuvan University, Kathmandu, Nepal, and Dr. Harka B. Gurung of the
  Nepalese National Planning Commission for their advice and assistance, and H.
  M. G. of Nepal for generously giving me permission to undertake research in
  Limi. I would, finally, like to express my gratitude to the various colleagues who
  have read drafts of this article and offered helpful comments.


  Benedict, B. (1972). Social regulation of fertility. In Hamson, G. A., and Boyce, A. J.
  	(eds.), The Structure of Human Populations, Clarendon Press, Oxford, London.
  Davis, K., and Blake, J. (1956). Social structure and fertility: An analytic framework.
  	Economic Development and Cultural Change 4: 211-235.
  Goldstein, M. C. (1971). Stratificatlon, polyandry, and family structure in Central Tibet.
  	Southwestern Journal of Anthropology 27(1): 64-74.
  Goldstein, M. C. (1974). Tibetan speaking agro-pastoralists of Limi: A cultural ecological
  	overview of high altitude adaptation in the Northwest Himalaya. Objets et Mondes
  	14(4): 259-268.
  Goldstein, M. C. (1975). Report on Limi Panchayat, Humla District, Karnali Zone. Contri-
  	butions to Nepalese Studies 2(2): 89-101,
  Muldock. G. (1949). Social Stnicture. Macmillan, New York.
  Nag, M. (1962). Factors Affecting Human Fertility in Non-industrial Societies: A Cross-
  	Cultural Study. Yale University Publications in Anthropology, No. 66, Yale Universi-
  	ty Press, New Haven.