"Selective Female Infanticide as Partial Explanation for the Dearth of Women in Viking Age Scandinavia," pp. 205-221 in Violence and Society in the Early Medieval West, edited by Guy Halsall. Woodbridge, United Kingdom: Boydell Press, 1998; paperback 2002. more

in Violenceand Sociery the EarlyMedieval'West Edited by Gry Halsall THE BOYDELLPRESS @ Contributors1998 All RightsRacrucd. Exceptaspermitred under curent legislation no part of this work may be photocopied,storedin a retrievd system, published,performedin public, adapted,broadcast, transmitted,recorded reproducedin any form or by any means, or without the prior permissionof the copyright owner First published1998 The Boydell Press,Woodbridge ISBN0 85rr57r30 The Boydell Press an imprint of Boydell & Brewer Ltd is PO Box 9, \Toodbridge, Suffolk IPl2 3DE UK and of Boydell & Brewer Inc. PO Box 41026, Rochester, 14604-4126, USA l.IY A cetdoguc recordfor this tide is evailablc from the British Libnry Data Libraryof Congrcss Catdoging-in-Publication / Violenccandsociety the earlymedierdr0flest editedby in Guy Hdsdl. P. cm. bibliognphicalrefcrences. Includes alk. ISBN 0-851l5-7lH (handcover: paper) !?'estcrn History' 2. Socid conflictl. Violence Europe, history- Mcdieval,50G-t500. Vestem- History. 3. Social Europe, I. Halsall,Guy. HN380.29V579 1997 97-20526 306.874 dc2t This publication is printed on acid-freepaper Typesetby Cruithne Press, 197 Great WesternRoad, Glasgow Printed in Great Britain bY St EdmundsburyPress, Bury St Edmunds,Suffolk Cuarrsn 11 Selective female infanticideas purtial explanution the for dearthof womenin Viking AgeScan dinavia' Nancy L. Wicker you are with child. If you should bear a girl, it shall be exposed, but if a boy, then it shall be raised.' Forsteinn toJ6fridr in Gunnl.augssaga This thirteenth-century Icelandic saga gives an example of ,exposure', or the abandonment of an unwanted child. Though the saga was written well within the christian period, Porsteinn's directive to his wife J6frf6r contradicts christian medieval teachings that dictate raising all children an-d perhaps recalls ninth-century norrns. From his declaration we may infer that men in this society had the right to make such decisions and that women were compelled to accept their judgement. His apparently cavalier attitude also suggests that female offspring may have routinely been disposed of in this manner. The action advised by Forsteinn, exposure, represents a category of violence rarely documented in Viking culture. Public violence in warfare and plunder was exalted and l. I have presented some of this material at the lgg2 ann '"r meeting of the Society for American Archaeology and at the l9g3 Scottish Archaeological Forum meeting on 'Death and the Supernatural.' Proceedings of the scottish Archaeorogical Forum paper will be published. 'So 206 Wicker mentioned frequently in the sagas and in history tending to form our views of violence in this society, while other forms of more private violence at home are little known. The Viking world has been considered one of powerfirl men, and we 'Viking women'. may question whether there even is such a concePt as 'Viking' apparently refers Though the erymology is undear, the word specifically to the seafaring men who pillaged and established colonies throughout much of Europe and beyond from 800 to ll50 $esch 1991: l). We know much less about women and children of the Viking period than about men because they were not similarly commemorated in life or in death in the extant histories and'sagas that we read or in the burial remains that we find. to any Here, female infanticide will be considered as contributing explanation of the scarcity of archaeological remains of women in some Scandinavian regions from the late Iron Age, c. AD 600, through the end of the saga period, c. AD 1300. Infanticide is recorded in Scandinasources during and shortly after the vian literary and protehistorical period of conversion to Christianity, around AD 1000, but the practice is difficult to trace in the archaeological record. Literary and historical scholars have noted infanticide in the form of exposure of infants, and archaeologists have detected a dearth of adult female burials in ttre late Iron Age, but these disparate studies have yet to be correlated with indirect archaeological evidence such as finds of scattered infant bones that may indicate exposure of infants. TTnwnrrrex EVIDENcE Though possibly repugnant to us, infanticide has been widespread throughout history and prehistory (Langer 1974; Ehrenberg 1989). Explanations for its occurrence revolve around tradeoffs pertaining to demography, ecology, and culture (Dickeman 1975; Hrdy 1992; Scrimshaw 1984). Elimination of infants of either so<, especially deformed or sickly individuals, may occur, but widespread parental favour for male children has also led to selectivefemale infanticide (Williamson 1976). Methods of infanticide vary. It often does not take the form of the cold-blooded murder of babies, occurring instead as the intentional abandonment or 'exposure' of undesired children or as the uncalculated preferential treatment, of the favoured sex and neglect of the other sex. Exposure was perhaps the most prevalent form in ancient Rome and Greece as well as Europe, but in other parts of the world drowning, suffocation, and strangulation were also customary (Lee l98l: 163). 'Deferred' infanticide by withholding adequate food and other re sources from unwanted children, which is less actively violent" also has been widespread (Johansson1984; Miller f98f). While not explicitly insinuating infanticide, Elisabeth Iregren suggeststhat withholding breast- FemaletnfanMe and the dearthof vlh@Age uomen 207 feeding may account for the unusually high mortality of infants from three to six months of age at Visterhus inJiimtland, Sweden(1988: 2225). Infanticide in all its manifestations has had a powerfrrl effect as a conscious or unconscious mechanism of population contol and the maintenance of social power. A cross-cultural examination of infanticide reveals a polarity between societies where infanticide is condoned, and those wherein it is condemned. Historical e>ramples the first type are known from ancient of Greece and Rome where infanticide may have been an acceptable method to control inheritance and dowry responsibilities (Bennett 1923; Golden 1990: 179; Harris 1982). A letter from Alexandria in the first century BC closely resembles the words of. Gunnlaugssaga in the intre duction to this chapter: 'If - good luck to you - you bear offsprin& and it is a male, let it live; if it is a female, exposeit' (t eftowitz & Fant 1982: 11f). As this selection informs us, newborns rarely were murdered with direct violence; instead, they were abandoned 'on garbage heaps or in public places' (Anderson & Zinsser 1988: 30; Bennett 1923). Following the socalled 'law of Romulus', many Roman families were unwilling to keep more than one daughter though they were obliged to raise all sons (Pomeroy 1975: 164). In Roman society, the paterfarnilias performed a ritual to recognize the newborn as a member of the family. He had the prerogative of deciding which children would be allowed to live, just as in Scandinavia, and if he chose not to recognize the infant, he ordered its exposure (Gardner 1986: 153-58; Pomeroy 1975). Since it was not yet a member of the family, the action was not considered murder. While exposure was not censured, it was carried out discreetly. Pomeroy asserts that ancient Greek and Roman court records and citizenship lists reveal only that daughters were much scarcer than sons, but never refer directly to exposure. As would be expected with a dearth of daughters, it seemsthat adult women are under-represented in the archaeological record in both Greece and Rome (Anderson & Zinsser 1988: 303l). In Christianized early medieval Europe, attitudes shifted toward vigorous social rejection of infanticide. Infanticide has been suggestedin tax and censrxrrecords (Coleman 1976, though problematic) and in clerical exhortations against infanticide (Sawyer 1987: 84). The vehemence with which clergy and law-makers attempted to eradicate the practice reveals their repugnance at this remnant from the pagan past. Rather than the father's preroga.tive, infanticide becomes more and more the desperate recource of an unmarried mother forced to kill her child singlehandedly to avoid social pressure upon herself and the child. Roslund (1990: 290) suggeststhis scenario to explain an infant skeleton discovered under the floor of an eleventh- or twelfth<entury house in Lund, Sweden. The tide turns from an action supported by communal values to one vilified by Christian societies (see Wrightson 1982 for dis. 208 Wicker cussion of this shift in attitudes). Christian authorities developed ways of attempting to deal with exposed children. These historical European instances of infanticide emanate from societies thoroughly conditioned by Christian condemnation of the practice. The Scandinavian written sources reflect a hybrid world where a veneer of moralizing teachings of the church met long-standing local tradition head-on. The topic of female infanticide in Scandinavia during and shortly after the period of conversion, around AD 1000, has been discussed by Carol Clover (f988), Juha Pentikiiinen (1968; 1990), and, most recently, byJennyJochens (1995: 85-93). Clover calls attention to infanticide in eleven Old Norse sources, including ,references to laws and customs as well as narratives describing specific incidents. The most straightforward Scandinavian accounts of infanticide, usually in the form of exposure, occur in Icelandic sagas. One such example (Gunnla.ugs saga, 3) is cited at the beginning of this chapter. Another example relates that 'Now it will be told that Earl Bjartmarr's daughter bore a baby girl, and it seemed advisable to most that it be exposed' (Heruarar saga, 3; Clover 1988: 158). Apparently exposure was considered appropriate simply because the child was a girl, not. because it was deformed or sickly. We have assumed that exposure is a method of infanticide, thus resulting in death for the exposed newborns. Boswell (1984; 1988) maintained that exposure in the sagas does not amount to infanticide and that exposed infants were meant to be found. Parents who exposed their offspring may have hoped for a propitious outcome for them. However, one must keep in mind that exposure in the cold harsh climate of the north could be deadly to a newborn much sooner than in Christian southern Europe, where the climate is kinder and the church's attitude was more pervasive. Each of the children intended to be exposed in the sagas was either prevented from being exposed, or discovered and saved before being harmed. In four instances, boys were put out to die but were discovered; these exposures clearly were considered unusual even within the saga context explicitly because the children were healthy males. Clover (1988: 150-58) contends that these saga examples were exceptional and that the discovery and adoption of foundling infants in Scandinavia must not have been common. These incidents were included only because the outcome was extraordinary and the child crucial to the story. Such stories may have been influenced by hero legends or the Oedipus folklore motif (Pentikiinen 1968: 69-73; Aarne 1964: no. 931). Children of special portent, such as Romulus and Remus, Oedipus, Moses, andJesus, have often been the focus of infanticidal intent. Clover (1988: l5l) also counters Boswell's thesis by pointing our rhar medieval law-codes, such as the Swedish Gutalagen, written down in the 1350s, included prohibitions against infanticide. Gutalagm (ch. 2) states: 'Every child born in our country shall be raised and not gotten rid of Femah infanticid.e and the dearth of Viking Age uomen 209 (Thomson 1960: 6). The practice of exposure was apparently common enough that laws against it became necessary in the Christian period. Pentikiiinen sees criminalization of exposure of infants as an indication of growing Christianization. The Icelandic governing body, ot thing, specifically allowed the heathen practice of the exposure of children during the conversion period, according to Ari Thorgilsson's Isl.endi.nga' b6k (ch.7; Flovstad 1956: 348). Its tacit acceptance underscores its existence. According to Icelandic sagas, a child could be abandoned until a name-giving rite was carried out: after this ceremony death by exposure was considered murder even in the pre-Christian period (Pentikiiinen 1990: 79). With Christianization, exPosure even before a child was named was criminalized. In the saga examples, perhaps the overlay of Christianity has affected the acceptability of such violence, since various people try to prevent exposure and save the child. The level of direct violence in these cases of intended infanticide is usually low. Normally the father orders the mother to abandon the child. Though expected to obey his command, she was not required to carry out the deed herself. Either Parent usually orders someone inconsaga sequential to the story to carry out the dastardly deed (e.g. Fi.nnboga out without specif2) or, more commonly, the child is ordered borne ing by whom (e.g. Gunnlaugs saga). Passagesin some sagas suggest that violence or the threat of violence may be used to dominate and regulate women in this society. Although usually coerced into submission, occasionally women seek actively to redress the situation and resist their husbands' orders. ln Fi.nnbogaSaga, a mother's objection is overmled by her fear of her husband, but J6fridr in Gunnlaugs saga defies Forsteinn and arranges to save the child. In at least one case, a woman could use the threat of exposure to control arrother woman's behaviour, as in Vatrxd,eln saga, in which the wife of a man who fathered an illegitimate child from an extramarital relationship orders the child exposed. Besides indirect infanticide by abandonment, other more violent methods are attested in the literature. ln Hardar saga oh Holrnunia (ch. 8), orders were given for a child to be thrown into a river. An Arab traveller, Al Tartushi, reported that infanticide was performed at Danish Hedeby, where unwanted infants were thrown into the sea to save the costs of bringing them up (Birkeland 1954: 104). A Christian apologist in Refidnln. saga (ch. 7) indirectly substantiates the custom of ritual infanticide by his report of the assembly's decision not to resort to sacri' ficing infans and elderly persons to propitiate the weather (Clover 1988: r52). Old Norse literary sources also mention fewer females than should be the case according to natural sex ratios. A suspicious preponderance of male children and a lack of female children, perhaps indirectly reflecting selective female infanticide, has been noted by Clover (1988: 167- 210 'lVicker 68) in lists of household membership in the medieval lcelandic Land, ruim.ab6k,a book detailing the tenth-century settlement of Iceland. The lists of household membership show that there were not as many girls and women as would be expected. Clover estimates that sons usually outnumber daughters at a ratio of four or five to one, occasionally even nine to one, perhaps indirecdy reflecting the effects of female infanticide. Swedish Upplandic runestones, as counted by AnneSofie Griislund, display similar ratios of sons to daughters. Griislund (1989: 233-40> suggests that female infanticide may account for this scarcity of daughters. Some scholars have attempted to discount evidence of infanticide by explaining that women and girls only seem to be lacking because they were not important enough to be mentioned as often. In either case, 'hidden' practice contributes to the relative invisibility of women. But even slaves were enumerated in Landruimabdk (I{arra"s 1988: 80), presumably because of their economic significance. It would seem logical, therefore, that each girl should also be noted due to the future negative economic impact that her dowry,, the woman's inheritance which was handed over by her father at marriage, would represent (Frank 1973: 475-76). While the scarcity of women and girls in written sources is not conclusive proof of infanticide, this testimony supports the proposition of female infanticide when considered alongside other evidence. Tne ARcHAEoLocIcAL EVTDENcE Archaeological evidence that might demonstrate the existence of infanticide is almost all indirect; the most perv-asivetestimony is the dearth of women's remains, as we find for Iron Age Scandinavia as well as for Merovingian Gaul (Halsall 1996: 15-16). John Boswell has called the analysis of sex ratios from cemetery remains 'a particularly treacherous methodology' (1988: 44, n. 107). While archaeological data have limitations, the source materials of historical demography, such as taxation lists, court lists, and ecclesiastical records, have their own weaknesses too (Welinder 1979: 33). Although estimating population from archaeological remains is fraught with difficulties, it should not be overlooked. Before discussing the scarcity of women's graves in Scandinavia, let us first examine how the sex or gender of a burial is determined. Idcntifrcationof sexor gmdcr in graaes There are two major methods to determine the sex or gender of a buried person: by analysis of skeletal remains, and by investigation of gravegoods. For Viking finds, both means are used because often the preservation of human remains is not satisfactory to permit conclusions on the basis of osteological analysis alone, due to cremation or soil con- Female infanticide and the dearth of Viking Age uornen 211 ditions affecting preservation. However, archaeologists are fortunate that many Scandinavians of the Viking period buried their dead with grave-goods. Vikings practised both inhumation and cremation, but even cremation graves with only fragments of burnt bone may contain grave-goods that may allow gender identification. The gender of the dead is often ascertained by grave-goods when skeletal remains are inadequate. Binford (1971: 22) proposed that the 'invisible most important way of displaying gender to the socalled society' of burial was by the objects in the grave: clothes, jewellery, other personal equipment and characteristic tools. However, it is debatable exactly which gravegoods can be used reliably as indicators of gender and, indeed, the whole basis for our assumptions of the gender specificity of artefact correlates needs to be re-examined (Brush 1988). In Scandinavia, weaPons and certain tools buried with skeletons normally indicate males, and jewellery and domestic implements signiff females, but it is diflicult to make these assignments with certainty. The usual practice has been to assume that a grave is male if, for instance, as few as one weapon was found, or female if five or more beads were found (Solberg 1985: 63). The numbers of objects decided upon may be arbitrary, but regardless of the number chosen as diagnostic, the artefacts used for gender determination normally are jewellery and weapons (Henderson 1989). Often there is substantial agreement between results from sexing graves by grave,goods and by analysis of skeletal material, but there is not complete accord (Henderson 1989; Pader 1981). A study of all unburnt Danish Iron Age burials (Sellevold et aI. 1984; Sellevold 1985) compared results arrived at by skeletal analysis to results indicated from gravegoods and determined that for the Viking Period there was a strong correlation between the sex of the deceased and artefacts buried with them. The objects that could be used to distinguish men's Sraves consisted of weapons such as swords and spears; axes; riding equipment including stirmps, bits, and spurs; blacksmith's tools such as shears, hammers, tongs, and files; and penannular brooches. Women's graves can be distinguished by the inclusion ofjewellery including paired oval brooches, trefoil buckles, disc brooches, artn rings, and necklaces; jewel boxes or caskets; and spindle whorls. Objects common to both sexes included buckles, combs, day pots, wooden vessels,knives, whetstones, coins, and, contrary to common assumptions, beads, sewing needles, and various jewellery types (Sellevold 1985: 67). Even the common assumption that weapons necessarily indicate males needs to be reassessed in light of a few problematic graves such as at Gerdrup in Denmark (Hemmendorff 1985) and Aurland in Norway (Dommasnes 1982: 77) in which women, as determined by skeletal analysis, have been found with a sword and arrowheads, respectively. IJltimately, however, sexing graves by either grave-goods or skeletal analysis indicates that women were scarce. 2t2 Indi.recteai.dence infantfuidc:thedparthoffemal.e of graaes Vicker A relative shortage of adult female mortuary remains compared to the e:rpected sex ratio of nearly 1: I has been noted in many regions of Scandinavia for the late Iron Age. Norway's population seems to diverge most markedly from average sex ratios. Dommasnes (f979; 1982; 1991) found a much smaller representation of women in studies of burials in four regions of the country. Ttre women's share of graves identifiable by gender in the four areas of Sogn, Gloppen, Nordland, and Upper Telemark varied from only 6To to 32Vo. Dommasnes (1979: 99-100) found ratios of eight males to one female in Sogn in the seventh century and six to one in the elghth century. The ratios are typical of graves throughout most of Norway in that period. For instance, Fllsn HOigard Hofseth (1988: fig. tf) found that women represented only from 8To to lSVo of the late Iron Age graves in Hordaland. In another study, Trond LOken (1987) found three times as many male as female graves in Iron Age material from Ostfold and Vestfold in Norway. In Denmark, sex ratios from cemetery analysis also are skewed toward males. The study of all unburnt Danish Iron Age skeletal remains found during the previous one hundred and fifty years identified 158 individuals of the Viking period for which sex could be determined by skeletal analysis (Sellevold et aI. 1984). Of these, 85 were found to be males and 73 females. The nurnbers represented are small and reflect, quite a sampling problem in Denmark where preservation is poor, but notably fewer women than men were identified and the sex imbalance is even more pronounced in earlier Iron Age material. In addition, just across the Danish4errnan border at Viking Hedeby, 62Vo of adult dead (47 individuals) that could be sexed skeletally were men and only 38Vo (29 individuals) women (Schaefer 1963). For Sweden there has been no country-wide re-evaluation of Iron Age skeletal material as completed for Denmark and in progress for Norway (Sellevold & Nass 1987), though there is a project underway for the medieval period (Iregren 1988: 25). The situation appqus to differ with a marked qualitative rather than quantitative difference between women's and men's graves. Studies of Swedish material have concentrated on extraordinary sites such as boat graves at Valsgiirde, as well as the large number of burials at Birka dating to Viking times. At Valsgiirde, men were inhumed in chamber graves and boat graves, but women were cremated (Arwidsson 1942;1954). At Birka where more than 2,000 grave mounds are visible, sex has been determined for only 415 burials. Griis. lund (1980) reports *rat women's graves there actually outnumber men's, representing 58% of the inhumations (308 burials) ar;id 61Vo of the cremations (107 burials). However, women were buried in the generally richer chamber gmves less frequenny $4Vo) tlnn men, so there was at least a qualitative differentiation between women'$ and men's graves. Fenale infanticide and the dearth' of Viking 'lge uorten 213 Women also made up most (687o) of those interred without coffins. Rather than indicating a preponderance of women at Birka, Griislund has suggested that the greater number of women's graves there may merely indicate that their graves are easier to identiff because of their contents, especially jewellery. However, Birka is anomalous; the trading community there should not be considered representative for the Viking period as a whole because of its unusual wealth and early missionary activity; the relatively large number of women's remains found at Birka might be explained by the missionaries' success. In her analyses of Norwegian material, Dommasnes (1982: 73) a* sumed that there was a l:1 ratio of men to l{omen, but PerhaPs that was not so. The sex ratio from cemetery analysis could be skewed if a portion of the population died elsewhere, away from home (Ehrenberg 1989: 127). One might expect that many men of Viking-Age Scandinavia died in foreign lands (Griislund 1989: 236-37), and at least some such deaths are memorialized on runestones commemorating men, listing where they travelled and who they fought (Morris, above: 149, 152). Warfare and migration could have taken such a toll on men that their remains would be scarce in'cemeteries at home (perhaps such as at,Birka). However, in many Scandinavian regions, men are not lacking: women are. Divale and Harris have hypothesized that preferential female infanticide compensates for the loss of adult males due to extra deaths in warfare (Divale 1970; Divale & Harris 1976). Such a functionalist explanation could explain the mirroring effects of public and private violence to regulate Viking society, a population in which heavy male outward migration and warfare might have led to an overabundance of women if not for the levelling effect of female infanticide at home. Perhaps because infanticide is so distasteful to us, some scholars have attempted to discount the dearth of women's remains in Scandinavia by explaining that women only seem to be lacking because they were not memorialized as often with large grave mounds or visible stone settings, so their graves go unnoticed. Dommasnes (1982), for instance' assumes she has not dealt with a representative sample of the Iron Age population. Women may have been given a different, less ostentatious, burial rite, as at Birka and Valsgiirde. Yet it is also possible that men actually outnumbered women due to selective female infanticide or other factors. We may be witnessing the results of preferential female infanticide compounded by the relative invisibility of low status female graves. Mortuary remaitu of infants and child'ren Very little direct archaeological evidence of infanticide exists' as the practice is rarely traceable in the physical record. Infant mortality was high in prehistory and the Middle Ages, often around 50Vo for Scandinavia (such as Welinder 1979: 83), so infant bones do not necessarily in- 214 Vicker dicate infanticide as the cause of death. The frustrations in locating adult female graves. wane compared to the difficulties of tracing children and infants in the archaeological record. Throughout much of pre history, children's mortuary remains apparently were disposed of in an archaeologically invisible manner, as at Udby on Sjalland in Denmark, where no children under age five were discovered in a Roman Iron Age cemetery (Ethelberg 1989: 7; cp. Morris 1987: 62). Iron Age children's graves are rare in Scandinavia. Gender- and age specific gravegoods are often all that remain since children's and specifically infants' bones rnay disintegrate easily depending upon soil conditions. The unossi{ied, cartilaginous skeletal material of infants may be totally destroyed during cremation; all that remains may be the unerupted tooth crowns and the densest temporal area of the skull (Henderson 1989; McKinley 1989: 242). Such scanty remains from earlier excavations may often have been missed altogether or discarded as apparent animal bones. Even when bones are preserved, they are fragmentary and have not developed the characteristic sexual dimorphism of adult specimens, so they rarely can be sexed osteologically (Lille hammer 1986: 12; Welinder 1989: 59). The only remaining method for assigning gender to them is according to gravegoods, with all the attendant problems of this approach for adult graves. Some Iron Age children's graves are very wealthy, as at Luistari in Finland (Irehtosalo-Hilander 1982:. 44-46), Birka in Lake Md.laren in Sweden (Griislund 1973; 1980), and Store Ihre on Gotland, where a young boy was buried with an adult-sized long-sword and a horse (Stenberger 1961). A newborn cremated at Mulde in Frdjel parish on Godand was buried in adult fashion but accompanied byjewellery sized to fit an infant (Lindquist 1989). Many children were not given an adult-type burial; some may have been deposited without any funerary ritual (Lille hammer 1986: l3). In many cultures, children are not considered fi.rlly human until some ctrlturally recognized threshold has been passed, such as when the child walks, talks, or is named (Williamson 1978: 64). Saga evidence reveals that children could be exposed until the namegiving rite was carried out, and those who died before this formality might not be considered fully-fledged humans deserving burial (Pentikiiinen 1968: 75). Occa"ional anomalous treatments of infant remains from diverse settings have been identified as the results of infanticide. Phoenician Carthage, well-known from Biblical and other references, is a notable instance where extensive infanticide is attested by archaeological as well as literary evidence (Brown 1986; Stager & Wolff 1984). Other examples of possible infanticide are recognized only through archaeological remains. Skeletal remains of numerous newborns found in a Late Roman/Early Byzantine sewer in Israel have been identified as the victims of infanticide (Smith & Kahila f992). Robbins (1977) has Female infanticideand,the dearthof VihingAgeuomen 215 suggested ttrat ritual infanticide occurred at the Incinerator site of the Fort Ancient culture in North America, where some infant bones were found in isolated burials unlike other infans buried in household cemeteries with diverse age groups. According to Robbins, these individuals exhibit cranial depression fractures consistent with being violendy thrown into pits as sacrifices. Mays (1993) deduces the practice of infanticide in Roman Britain by demonstrating that the age at death of of newborn infant remains is not consistent with the distribution modern stillbirths and natural deaths. in Scandinavia is Direct archaeological evidence of infanticide (1981) has proposed that infant bones in stone limited. Malin Lindquist packings of cairns may be remains of infanticide victims. Inger Saelebakke (1986: 24) suggests that apparently haphazard bone scatters in middens may be interpreted as remains of exposed infants. At a Norse site at Buckquoy, Orkney, infant bones dispersed through middens may testif to infanticide (Ritchie 7976-1977: 188, 220-21). The scattered remains initially were mistaken for animal bones upon discovery only during osteological analysis were they identified as human. Other e:ramples from such contexts indude five infant skeletons strewn through a garbage pit excavated at S6rby Skola, Giirdsl6s4 Sweden (Sj& berg & Marnung 1976). Many middens as well as other bone caches from earlier excavations were not investigated as thorougtrly as these examples, and one wonders how many infant bones have thus escaped discovery. Archaeologists should carefully investigate bone scattersr particularly from middens, to search for infant remains, as suggested by Igrje (1989: 70-74), who cites an example of infant bones from Fr6s6n inJimtland, Sweden, mistakenly identified as animal bones in 1910. Infant bones also have been found in wells and bogs, such as at R6ekillorna, Skine, where the skull of a small child was found in a well (Stjernquist 1987) and Bo in Hi komrnune, Rogaland, where four newborn skeletons were found in a bog (Haavaldsen 1989). Discoveries of bones in these watery contexts often have been interpreted as evidence of sacrificial infanticide, but also concur with Al Tartushi's description of infanticide by drowning at Hedeby (Birkeland 1954). Children's burials become more numerous in Scandinavian medieval cemeteries than in the pagan grave.fields, perhaps in part because the church's opposition to infanticide limited the practice somewhat (Sawyer 1987: 4). As an explanation for the increased number of older children's graves, Ole Benedictow (1985: 23) has pointed out that'the prohibition of exposure of neonatals . . . possibly might have led to an increase in the death rates of somewhat older children through malign neglect.' An example of this cited earlier may be demonstrated by the cemetery remains at Viisterhus (see above) where Elisabeth Iregren (1988: 22-25) attributes numerous remains of infants from age three through six months to early cessation of breast-feeding. However, the most import- 216 Wicker ant reason why more child burials of the Christian period are found is that according to medieval laws, the baptized were required to be buried in consecrated ground. The law also exduded the unbaptized from burial in Christian cemeteries, so dying infants usually were baptized quickly so that they could receive Christian rites (Karlsson 1988; O'Connor 1991; Sigurdard6ttir f983). At the medieval site of Helgeandsholmen in Stockholm, some skeletons of newborns were found just outside tlre wall surrounding the cemetery (Jacobzon & Sj<iberg 1982:12223). That they were not interred in the cemetery suggests they had not been baptized. If they were not baptized, we may hypothesize that they did not die by natural causes; perhaps they were victims of infanticide, as Roslund (1990: 286-90) proposed for the newborn (above: 207) that was buried under the day floor of a house in Lund, Sweden, dating to the early Christian period there. ConcrusroN of these scattered archaeological traces of infants The interpretation who were not buried with an adult-t1pe burial rite is problematic. Even if we can demonstrate that infanticide was practised, we have only indirect grounds for discussing the existence of preferential female infanticide. However, we should not underestimate their worth in detecting this practice. Carol Clover (1988: 165) has called archaeological evidence 'lesser forms of evidence'. I would suggest that the testimony one of the of hundreds of graves seen within the context of the literary and historical evidence is a potent source of information. Female infanticide could account for the dearth of women in literature and in cemetery remains, inand all the data serve as supporting evidence in an interdisciplinary vestigation. Infanticide is so pervasive throughout history that it should not be surprising that it occurred in Scandinavia. Examining this hidden practice of violence in the Iron Age and Middle Ages may complement our view of the outward violence in these societies. Viking men were re viled for their pillage of Christian Europe, but women most directly suffered the effects of the ruthless regulatory mechanism of infanticide, since many Viking girls apparently were not allowed to live to become 'Vikingwomen'. BnuocRApHv Pri.mary Sources Inndrfimabflh. Ari Thorgilsson, islnrdingaDd&: Benediksson (ed.), islmdingab1k. J. islmzhfomrit l. I (Reykjavik,1968). saga,isbnzk Fomri.t 14 (1959), Finnbogasaga:ln !. Halld6rsson (ed.), Kjalncsinga pp.251-340. Fenale lnfantic'tde ond the dearthof VihlngAgeuomen 2r7 Gunnlaugs saga Omutungu: S. Nordal and G. J6nsson (eds), Boryfirninga sogur. Isbnzhfomrit 3 (Reykjavik,1938),pp. 5l-107; G. Jones (trans.),Erih and Rzd, (London, 1961,reprint 1980),pp. l7l-217. ard Otlur lczla,ndic Sagas Gu.tala.gm:C. J. Schlyter (ed.), Corptu iuris nuogothmum ontigui. Sanling Af Sunriges Gaml.a Lagar 7 (Lund, 1852). Harlar saga oh Hdlnvefa: S. Hast (ed.), Editionar ArnamagnacarncA 6. (Copenhagen, 1960); A. Boucher (trans.), Tlu Saga of Hord and tlw HolrvDrcllerc (Rey$avik, 1983). Hennrar saga.Heidrehs saga(Hennrar sagaok Heihrehs horutngs): Helgason (ed.), J. tumfund. til Udgtuebe GammelNmdishLittzmtur 48 (Copenhagen, 1924, rpt. af r976). I-a.rd.ruimahih: Benediktsson (ed.), Islad.ingaMh. Lonfunmatdh. Islznz.k fomrit I J. (Reykjavik, 1968); H. Pdlssonand P. Edwards (trans.), Tlu Boohof Settlznents: (Winnipeg, I 972). Land.ruina.hik Rcyhdetasagaok Vtga^Sktitu; Sigfilsson (ed.), Ljdxntninga nga. Islznz.kfomrit l0 B. (Rey\iavik, 1940),pp. 149-24; W. H. Vogt (rans.), 'Die Geschichtevon den Iruten aus dem Rauchtal', in F. Niedner and G. Neckel (eds), Thulc: Altnord.isclv Dichtung unil Prosavol. 11 (DrisseldorFKdln, 1921, reprint 19il), pp.297-372. Vatrud.ela, nga:, E. O. Svinsson (ed.), Vatradnla.saga.Istenz.k fomnr 8 (Rey$avik, (Princeton, 1944). 1939);G. Jones (rans.), Tlw Vatrudalen'Saga Secondarybibliography Aarne, A., 1964. The fW of tlu Folknle. A Classifitation and Bibliograplry, trans. and enlarged by S. Thompson 2nd revised edition (FF Communications 184) (Helsinki). Anderson, B. S., & Zinsser,J. P., 1988. A History of Tluir Oun: Womenin Europ to from PrehLrtory thc Pramtvol. I (New York). Arwidsson, G., 1942. Valsgd,rdz Dio C,rdberfunde Vahgd,rdz, (Uppsala). 6. I uon 1954.Valsgd,rdz Die Grawndz vonValsgdrdz (Uppsala). 8. 2 Benedictow, O. J., 1985. 'The Milky Way in history: breast feeding antagonism between the sexes and infant mortality in medieval Norway', Scandhadan Journal of History10: 19-53. Bennet! H., 1923. 'The exposure of infants in ancient Rome', Classieol Juurnal l8: 341-51. Binford, L. R., 1971. 'Mortuary practices: their study and their potential', in J. A. Brown (ed,.), Appoaclw tn tlu Social Dimm.sionsof Morhnry Practic* (Memoirs of the Society for American Archaelogy 25) (Washington, DC), pp.6-29. Birkeland, H., 1954. Norderc Historie i Midndaklnen efizr ArabisheKilder (Norske VidenskapsakademisSkrifter) (Oslo). Boswell, J., 1984. 'Fapositio and obla,tio:the abandonment of children and the ancient and medieval family', Ancrimn Histmical Retiru 89: 10-33. 1988. Tlu Kindnars of Strangen: Tlu Abandnnmentof Children in Wstern (New York). Eumpfrom Inte Antiquity to tlu Rnlaissanaa Brown, S. S., 1986. 'Late CarttraginianChild Sacrificeand SacrificialMonuments in Their Mediterranean Context' (Ph.D. Dissertation, Indiana University). Brush, K., 1988. 'Gender and mortuary analysis in pagan AngloSaxon archaeology', Arclwohgi.ml Reoieufrom Cambridge 76-89. l: 2t8 'lVicker Clover, C. J., 1988. 'The politics of scarcity: notes on the sex ratio in early Scandinavia', Scand.irnvian Stud,i?s 147-88. 60: Coleman, E. K., 1976. 'Infanticide in the early Middle Ages', in S. M. Stuard (ed.), Womm in Medianl Socizty (Philadelphia), pp. 47 -70. Dickeman, M., 1975. 'Demographic consequencesof infanticide in man', Annual Rnian of Enlog and,Sytnwtia 6:107-37. Divale, W. T., 1970. 'An explanation for primitive warfare: population control and the significanceof primitive sex ratios', TIw Neu Sdplar2:173-79. Divale, W. T., & Harris, M., 1976. 'Population, warfare, and the male suprematist complex', Amcrban Anthropologist 8: 52l-38. 7 Dommasnes, L. H., 1979. 'Et gravmateriale fra yngre jernalder brukt til i belyse kvinners stilling', Vihing 1978,pp. 95-l 14. 1982. 'Late Iron Age in western Norway: female roles and ranls as de duced from an analysis of burial customs', Norungian Archazolagbalfuuieto 15:70-84. 1991. 'Women, kinship, and the basisof power in the Norwegian Viking Age', in R. Samson (ed.), SocialApproatlus to Viking Shrdiat (Glasgow), pp. 65-74. Ehrenbery M., 1989. Women Prehisnry(London). in Ethelbery P., 1989.'Skridbanken', Shalh2: 3-9. Frank, R., 1973. 'Marriage in twelfttr- and thirteenth<entury Iceland', Viator 4i 478-U. (Bloomington IN). Gardner,J., 1986. Wonen in RornanLa,ward Society Golden, M., 1990. Chi.ldren and Childhoodin Classical Athcns(Baltimore). Grislund, A.S., 1973.'Barn i Birka', Tm 15: 161-79. 1980. Birha IV. Tlu Burial Cutams. A Stttdyof tlw Cmws on Bjdrho. (Birka, Untersuchungen und Studien) (Stockholm). 1989. "'Gud hj?ilpenu viil hennessjil:" om runstenskvinnorna,deras roll vid kristnandet och deras ptats i familj och samhille ' , Tm 22: 22344. Haavaldsen,P., 1989. 'Pojken mi fdrast ut i kirret . . .', Fni Haug ok Heidni 12: 174-76. Halsall, G., 1996. 'Female status and power in early Merovingian central Austrasia the burial evidence', Eady Med.inal Eump 5: l-24. Harris, W. V., 1982. 'The theoretical possibility of extensive infanticide in the Graeco-Romanworld', ClassbalE arttd, 32: 114-16. Hausfater, G., & HrE, S. B. (eds), 1984. Infarxbifu: Comparutiu ard Ewlutimnry Penpccttua,(New York). (Hojbjerg), pp. Hemmendorff, O., 1985. 'Gravensbipersoner', inSka.Ihs Gestzbog 13-20. Henderson,J., 1989. 'PaganSaxon cemeteries:a study of the problems of sexing by gravegoods and bones',in Roberts,lce, & Bintliff(eds) 1989:77-83. Hoignrd Hofseth, E., 1988. 'Liten tue velter . . . problemer knyttet til manns' og kvinnegravenes i fordeling i Nord-Rogaland', Atxihhcl-samling II (AmS Skrifter, 12) (Stavanger), 5-38. pp. Hovstad,J., 1956. 'Barneutbering', in;. Granlund (ed.), KuhurhtttorishtLwihon vol, I (Malm6). for Nonlbh Medzltidfrdn Vihingatd till Refonnationstid HrE, S. B., 1992. 'Fitness tradeoffs in the history and evolution of delegated mothering with special reference to wet-nursing, abandonment, and infanticide', Ethohg and Sociobiobg 13: 409-42. Fentale infanttuide thefuarthof YfhfngAge and uomen 219 Iregren, E., 1988. 'Avbruten amning blev barnens d6d? - Ett forsok till tolkning av Vesterhusmatenalet, Populir Arhzolngi4: 22-25. 'Miinniskor i stadensutkanternrHelgandshusea Jacobzon,L., & Sj6gren,J., 1982. Helgeandslnlmst kyrkognrd berdttar om liv och d<id', in G. DaHbeck @d,.), 1000 d,r i Stochh.olrns (Stockholm), pp. 112-33. str6z Jesctr,J., 1991. Woncn in tlu Viking Age (Woodbridge). (Ithaca & London). in Socicty Jochens,J., 1995. Wornen Old Norse S. R., 1984. 'Deferred infanticide: excess female mortality during Johansson, childhood', in Hausfater& Hrdy (eds) 1984:463-85. Karlsson, I., 1988. 'Barnets Stiillning i Norden under Medeltiden - en analys med utgingspunkt fri.n arkeologiskt och osteologiskt material', Uppsats i medeltidsarkeologi Ak 200 vid Lunds Universitet (Lund). Karras, R. M., 1988. Slawy and Society Medi.anl Scandfuwaia(New Haven & in London). [,anger, W. L., 1974. 'Infanticide: a historical sulvey', Hitt rl of Chikllwotl Qurtcrty l: 353-65. La{e, R., 1989. 'Benfrnd urjdmliindska gravar', in O. Hemmendorff led.), Arhzo Iogi i frAU shogoch bygd.2. Jamdldzr+nedzltil.(Fornvirderen 24) (Ostersund), pp.61-77. I-ce, B., 1981. 'Female infanticide in China', Historbal refuctions. Rclbxions histotiqnt 8:163-77. Irfkowitz, M. R., & Fant, M. 8., 1982. Wonm's Life in Green and,Romc:A Source Boohin Trarulation (Baltimore). LehtosaloHilander, P.-L., 1982. Lui,snri III. A Burial4rowd Rtfl"rti"C the Finnbh Vihing Age Socufl (Suomen Muinaismuistoyhdistyksen Aikakauski{a, Finska Fornminnesforeningens Tidskrift 82: 3) (Helsinki). Lillehammer, G., 1986. 'Barna i Nordens forhistorie: droft metodegrunnlaget og kildenesberekraft', IQtirmn i Arheologi Norge2:3-21. i Lindquist, M., 1981. 'Mylingaroffer, utsatta barn eller forhistoriska barnbegravningar?', Gotlindsht arhio 53: 7 -12. arhtu 6L:241-42. 1989.'Barngravarvid Mulde i Friijel sn', GotUindsht kken, T., 1987. 'The correlation between the shape of grave monuments and sex in the Iron Age, based on material from @stfold and Vestfold', in R. Bertelsen, A. Lillehammea & J.-R. Nass (eds), Wen Tluy All Mm?: An Examina.tionof 9x Rolcsin PrehistoricSoc&ty(Ams Varia ff (Stavanger), pp. 53-63. McKinley, J., 1989. 'Spong Hill, AngloSaxon cremation cemetery', in Roberts, ke, & Binttiff(eds) 1989:241-48. Mays,S., 1993.'Infanticide in Roman Britain', Anti7"ttl6T: 883-88. Miller, B. D., 1981. Tlu Endangend Sx Neghct of Ferwlz ChiWm, in Rural North lzdra (Ithaca,N.Y.). Citystate (C,zrft Morris, I., 1987. Burial and Arcient Sociar: Tlu Riseof tlu C'reeh bridge). O'Connor, A., 1991. ChiLdMurdzressand Dcad Child. Traditiolrs (FF Communications,no. 249) (Helsinki). Pader, E.-J., 1981. SymbolisrySocial Rehtions and tlu Intzrprctation of Mmhnry Remairu (BAR International Series, 130) (Oxford). Pentikiiinen, J., 1968. Tlu Nudic fuad,ChiW Traditiu. (FT Communications, no. 202) (Helsinki). 220 - Wicker 1990. 'Child abandonment as an indicator of Christianization in the Nordic countries', in T. Ahlbiick (ed.), OId. None and. Finnish Religions anl. (Abo), pp. 72-91. CukicPlaceNanas Whnra, Wiaes,and Slattes:Worwn in Classical Pomeroy, S. B., 1975. Goddesses, Antiquig (New York). Ritchie, A., 1976-1977. 'Er<cavationof Pictish and Viking-age farmsteads at Buckquoy, Orkney', Proccedings tlu Sociztyof Anti.qtnrics of Scotland 108: of tl4-227. Robbins, L. M., 1977. 'The story of life revealedby the dead', in R. L. Blakely (ed.),BioculturalAdnptationin Prehistmic America(Athens GA), pp. 10-25. Roberts, C. A., Lee, F., & Bintliff, J. (eds), 1989. Burial Archncolog: Curmt (BAR British Series,211) (Oxford). Methods Dnelapmenrr and Roslund, M., 1990. 'Nittiotalets medeltidsarkeologioch det doda barnet', Fonr odnnzn 85:283-92. Saelebakke, 1986. 'Noen f6relobige inntrykk fra en middelalder - kirkegard i I., Tcinsberg', META 2-3: 2l -25. Sawyer,P., 1987. 'The processof ScandinavianChristianization in the tenth and ' eleventh centuries', in B. Sawyer,P. Sawyer,& I. Wood (eds), Thz Christian(Alingsns), pp. 68-87. izati.onof Scandi.na,aia Schaefer,U., 1963. Anthropologisclw t)on Haithabu (Die Unterrhung dzr Shzlette Ausgrabungenin Haithabu, 4) (Neumrinster). Scrimshaw,S. C. M., 1984. 'Infanticide in human populations: societal and individual concerns',in Hausfater& Hrdy (eds) 1984: 439-62. Sellevold, B. J., 1985. 'Knokler, oldsaker og kvinner: Ssisk antropolgi som metode til kunnskap om kvinner i middelalderen', in Kainruarbei.di. Nordzn fra aikingtiden ti.l reformasjorwn(Foredrag fra et nordisk kvinnehistorisk seminari Bergen3-7 August 1983)(Bergen),pp.63-77. Sellevold,B.J., Lund Hansen, U., &Jorgensen,J. B., L984.Iron AgeMan in Dm, mar&(Nordiske Fortidsminder 138)(Copenhagen). Sellevold, B. J., and Nass, J.-R., 1987. 'Iron age people of Norway', Noruegian Archazologiral Reuieu20: 46-50. Sigurdard6ttir, A., 1983. 'Ret er at en kvinde larer ham at dobe et barn: Om dib, konfirmation og fadderskabi Island i middelalderen', in S. Adalsteins d6ttir & H. Forlfksson (eds), Fdrdndringar i kainnors aillhqr undzr mcfultidzn (Rey$avik),pp. 4l-54. Sjoberg, M., & Marnung, B., 1976. Fomlnrnning-126. GraaJidtt, Aldre J'drnildcr, Sdrbl Skola" Ostcr Sdrb1, Gd,rdsl\sa sockeq Ohnd. (Riksantikvarieiimbetet Rapport, B44) (Stock'holm). Smith, P., & Kahila, G., 1992. 'Identification of infanticide in archaeological sltes: a casestudy from the late Roman-early Byzantine periods at Ashkelon, lsrzel' ,Journnl of Archaeologi.cal Sciencc 667-7 5. 19l. Solberg, 8., 1985. 'social status in the Merovingian and Viking periods in Norway from archaeological and historical sources', Noruegian Arclwzologiaal Reuiaa18: 6l-76. Stager,L. E., & Wolff, S. R., 1984. 'Child sacrifice at Carthage:religious rite or population control?',BiblicalArchaeologr, Jan.-Feb.1984,pp. 30-51. Stenberger,M., 1961. 'Das Griberfeld bei Ihre im Kirchspiel Hellvi auf Gotland. Der wikingerzeitliche Abschnitt', Acta Ardnzologica 32: l-134. { of Age and Female infanttutde thedearth Viking uornen 221 Sqjernquist, 8., 1987. 'spring-cults in Scandinavian prehisloryl, in-.T. Linders & G.-Nordquist (eds),-Gi7futo tlu Gods(Acta Universitatis Upsaliensis Boreas, 15) (Uppsala),pp. 149-57. Thomson, A., 1960. Bamhfitmingm: En rdttshistmishStzdie (Skrifter utgivna af Kungliga Humanistiska VetenskaFssamfundeti Lund, 58) (Lund). (Acta Ardraeologica Lundensia, Series Welinder, S., 1979. PrehistmicDernography in 8eMinore, S) (Lund). 1989. 'An experiment with the analysis of sex and gender of cremated bones',Tm22:29-41. Williamson, L., 1978. 'Infanticide: an anthropological analysis', in M. Kohl (ed-)' Infanticitz antl tlu VaIw of Li.fe(Buffalo), pp. 61-75. Williamson, N. E., 1976. Sorcor Daughtm: A Cross4ulnral Sutvel of Parmtal Pre (Beverly Hills). ferences Wrightson, K., 1982. 'Infanticide in European history', CrimhwlJruti.u Histmy gt l-20.