0utlaws of Surtshellir Cave: the Underground Economy of Viking Age lceland
    Guõmundur ?lafssonl, Thomas H. McGovernz and Kevin
    P.
    
    Smith3
    
    lrurRooucnolrt
    ln 2001 investigations by the National Museum of lceland and Brown University documented a suite of significant archa eologic al featu res and d eposits within the Surtshellir-Stefánshellir cave system, lceland's longest lava tube cornplex, located within the H allmundarh ra un lava field, in Hvítá rsíða, western lceland (see Fig. 1). This project was undertaken to establish the age and integrity of deposits present in the cave, following reports that tourists were removing bones from it as souvenirs (Ólafsson et at.20041. ln the 13th century text Landnámabók (Pálsson and Edwards 1972), the 14th century Harðar saga ok Hólmverja (Jónsson 1953), and the fictive 19th century Hellismanna saga (Árnason 1889) Surtshellir is identified as the base of an outlaw band that fortified the cave in the late 1()th century preying on the surrounding countryside until routed by a
    
    Fig. l. Location of Suttshellir in lceland.
    
    coalition of local chieftains. The presence of bone piles and a stone-walled structure have been anecdotally recorded since the 17th century and a single radiocarbon date was run on a Bos bone fragment in 1969 by the lcelandic author Halldór Laxness. These fragmentary records have been used in the past to suggest archaeological suppott for the historical narratives. However, the calibrated age of Laxness'date spans the entire Settlement and Commonwealth periods (AD 870-1264) and could reference activities undeftaken at any point in the early Middle Ages. Furthermore, in the absence of any detailed archaeological record of the cave's cultui'al features, no framework existed for protecting its archaeological resources' assessing the nature of the activities that once took
    
    place in Surtshellir, or comparing its deposits with those from other lcelandic sites associated with outlaw activity by historic reference or archaeological inferences. Surtshellir is one of thirty-three lcelandic caves with evidence of human occupation or activities from the time of lceland's initial Norse settlement (ca. AD 874) to the recent past. Seven are known within the Hallmundarhraun; two of these had received some archaeological attention prior to these investigations (Gestsson 1960; Ólafsson 2000). Fieldwork in Surtshellir therefore provided opportunities to determine whether the cave's archaeological deposits could be dated to the Settlement period, to inventory and characterize the archaeological features and deposits within it, possibly to gain new Ínsights into outlaws'material culture and ¡ -- aL-,--:-^^l uurrorru tnelr lfnpacls oll tf le lcglullal sLUllullly, ^-l+^ fu ^^^ the study of tribute data from the North Atlantic to caves in the archaeologieal record'
    
    1. National Museum of lceland.
    
    2. City University of New York. 3. Haffenreffer Museum of Anthropology, Brown University.
    
    395
    
    Three features were identified, mapped, and investigated at Surtshell¡r in 2001: an intact drystone dwelling (FiS. 2); a well-preserved, bone-rich midden beside it (Fig.3 and 5); and a massive fortification wall. All are located deep within the cave.
    The fortification wall, 13 meters long and standing, in places, to a height of 2.5 meters, was located
    
    '-1-\-/r ^. -È^^r^À1' 1
    
    nearly 100 meters into the cave, on the edge of an opening produced by roof collapse. More than 70 meters farther into the cave, a secondary passage opens 5 meters above the floor of the main tunnel and crosses it at an acute angle. The northern spur of this passage, known as Beinahellir ("bone cave"), is said once to have contained extensive deposits of animal bone, although none exist there today. Howeveç thirty meters into the southern branch ofth¡s upper passage stands a unique drystone structure with an associated midden, nearly 200 meters from the cave's entrance. The presence of structural remains and associated midden deposits in this passage, called Vígishellir ("fortress cave"), has been known since at leastthe late lTth century (Þórðarsson 1910; Laxness 1g49, t97t; Ólaf-
    
    ^s*,¡níîni
    
    'f'€x''4;i
    -ì
    
    anll:'
    
    rrÌì
    \\
    
    E viMoL,tur
    
    Fig. 3. Plan of stone-built structure and midden de-
    
    posits in Vígishellir.
    
    sson 1978; Hróarsson 1990; Jónsson and Hróarsson 1992), but had never been systematically documented until the 2001 investigations. The structure in Vígishellir is 7 m long and 3.b m wide (Fig. 2 and 3). lts 0.5-0.8 meter thick dry-stone walls stand 0.8-.l.0 meters high and appear to be undisturbed, suggesting that this structure, whích may have had a perishable superstructure, is substantially intact. Like many Viking Age halls it narrows towards the gables where it is 1.5 m wide. ïwo doorways, each 0.8-1.0 meters wide, located in the centers of the two long-walls, give access to. the structure and to the depths of the cave beyond. However, the structure itself is built across the passage, blocking access to the cave's darker
    çuçoùsù, ..,L^.^ç ^^ ouutLtuilot t uttul dl uupu5ll,5 vvt tEt ttu ^,J-l¡+i^^^l ^..1+,.,-l l^^^^:¿^ have been observed. ln the eastern gable a niche in
    I
    
    Fig. 2. }verview of the building with curved walls and a gable-end fireplace. (K. Snith).
    
    the wall forms a fireplace; a second niche ín the southwestern wall may have been used for storage. No traces of benches or sleeping alcoves are visible. ln the central and eastern portions of the structure a thin (<1 cm), compact, unstratifíed floor
    
    396
    
    r I
    deposit of dark gray ash with small fragments of calcined bone lies directly on the unmodified basalt cave floor.
    
    I ì
    
    MRrenrnl
    
    cULTURE
    
    Four jasper fire-starter spalls were recovered within the structure and on the pathway leading to it (Fl-3,5 on Fig.3). Trace element signatures for
    
    these objects were determined by Smith and Ronald V. Hancock through lnstrumental Neutron Activation Analyses undeftaken atthe Royal Canadian Military College's SL0WP0KE reactor facility in Kingston, 0ntario. Concentrations of manganese, calcium, and vanadium in these specimens conform to values from known lcelandic geological sources, but are distributed widely among them. While the Vígishellir samples cannot yet be linked directly to any specific sources through trace element analyses, variability among them clearly differs from that found in jasper fire-staner assemblages from two other Viking Age sites in the same region. For example, fire-starters and spalls from Háls 5, a 1Oth century iron-production complex located 20 kms from Vígishellir, cluster tightly with local jasper geochemistries. Nine fire-starter spalls from a 10'h century outlaw shelter in the cave Viðgelmir, less than 5 kms south of Vígishellir, match the chemical signature of jasper from Hestfjall, 50 kms to the southwest. Greater variability within the small suite of spalls from Vígishellir suggests that this cave's occupants traveled over greater distances or were drawn to the cave from a wider region.
    
    Fig. 4. The surface of the Vígishellir midden (K. Smith).
    
    1960s and Eggert Ólafsson's description from the
    
    Mtoorru
    To
    
    EXCAVATToNS
    
    late 17th century both implythatthis bone bed was once much larger in extent and that its uppermost surface was once covered with highly decomposed and fragmentary bone bits, Three layers were identified and sampled in the field: a light dry upper layer (Fig. a), a somewhat darker but still dry middle layer and a darker brown ash-enriched and water-soaked deposit of fine bone fragments lying directly upon the basalt floor of the cave. Bulk matrix samples taken from the excavated unit recovered a 100% sample of all bones, artifacts, and associated sediments present in the three layers. The bulk samples were dry sieved in the laboratory and sub-divided by screening through 114" 14 mm)and 1/16" (1 mm) mesh sieves. All sediment passing through the 1 mm screen was saved for micro-artifact analyses; all bones and bone fragments retained on the 4 mm and 1 mm mesh sieves were submitted to the NORSEC laboratories (City University of New York) for analysis.
    
    the east of the house is located the remains of a large, bone-rich midden. A 1.25 x 0.5 m test unit measuring was excavated near the center of this 3.7 x2 m deposit, sampling 10% of the midden. The excavated cieposit was 3-7 cm thick, yet according to the landowner was once 30-50 cm thick. Pockets of bone and ash adhering to the cave's eastern wall support this assertion and suggest that the deposit sampled in 2001 represents the basal remnant of a once more substantial midden. lt is worth noting that Laxness' brief mention of the midden from the
    
    DRrrrue
    Tho rianncil nan ho riatori hnih inriiro¡tir¡ thrnr¡nh ¡ vvPvùt!
    "v
    
    tephroc hronology a nd directly throu gh radio ca rbon dating. The lava field in which Surtshellir and Vígishellir formed, H allmundarhraun, rests a bove the Landnám tephra layer (Jóhannesson 1989),
    
    which has been independently dated to AD 871 +/- 2 (Grönvold etal. 1995). Two AMS radiocarbon dates
    
    397
    
    Table L Radiocarbon dates from Vígishellir cave, Hvítársída, lceland.
    
    AAH-7412 (top layer)
    1214x41 BP [AMS, Bos taurus bone fragment] One sigma (68.2% probability) calAD 770-B9O Two sigma (95.4% probability) calAD 680-950 Í680-900 (92.7%),920-950 (2.7%)l
    
    AAR-7413 (basal layer) 1197i36 BP [AMS, Bos taurus bone fragment] One sigma (68.2% probability) calAD 770-890 Two sigma (95.4% probability) calAD 710-9601710-750 (5.8%),760-900 (82.6%),920-960 (7.0%)l
    K-1435 (Laxness I969) 1010*100 BP [Standard, Bos taurus bone fragment] One sigma (68.2% probability) calAD 900-11601890-920 (5..4%),950-1160 (62.8%)J Two sigma (95.4% probability) calAD 7BO-12401780-792 (0.9%), 804-1223 (94.2%) 1233-1237 (0.3%)j
    
    (AAR-7412 and AAR-7413) run on collagen from cattle þones collected in 2001 imply a maxirnum age
    
    for the deposit of calAD 690-960 With the Landnám tephra providing a firm terminus post quem for the formation of the cave itself, the bone-bearing deposits tested in 2001 can be reasonably dated to the period AD 880-960 with a high degree of certainty, Although these ¡ruo AMS dates cannot be statistically separated, suggesting a rapid accumulation of the sampled deposits. The cow bone dated by Halldór Laxness (1971) produced a somewhat later calibrated age (Table 1 [K14351). While it is unknown whether Laxness' sample was recovered from the deposits sampled in 2001 or from higher levels of the same midden, since removed, that date overlaps both of the new AMS age ranges attwo standard deviations. Together, the three dates suppod a late 9ú-1()th century age that is generally consistent with the period of outlaw activity noted in medieval accounts.
    
    TapH
    
    or.¡
    
    orr¡rc 0
    
    BSERVATT oNS
    
    As it appears that this bone deposit accumulated over a fairly short period of time, its stratigraphic divisions are unlikely to represent different phases
    of occupation. Water percolation and fi'eeze/thaw cycling in the exposed bone bed most likely produced some of the size sorting evident in the collection - the wet basal layer is almost entirely conr-
    
    posed of tiny bone flakes and chips. While the current archaeofauna derives from a small excavation unit, it easily exceeds the NAB0 minimum sample
    
    size limits (300 mammal bones or 1000 fish bones) for basic comparability. Bones from Vígishellir are very fragmented and show signs of rather specialized weathering resulting from their exposure on the cave's floor. However, most fragments were quite robust and not excessively damaged bythe exfoliation and "mushiness" evident in bone exposed to weathering in open air or an acid soil matrix. The highly fragmented nature of the collection reflects extremely complete butchery and bone marrow and collagen extraction prior to deposition, with some flaking and spalling afterwards. Table 2 compares the fragment size distribution for the Vígishellir collection with an approximately contemporary collection from Hrísheimar in northern lceland near Lake M,ivatn (McGovern and Perdikaris 2002). The Hrísheimar context 003 collection, from a small excavation unit (2 x 2 m) that sampled midden filling a sunken-featured structure, was also sieved through 4 mm mesh. Three AMS dates on cattle bone collagen place this deposit in the late gth to 10'h century (calibrated 1 sigma ranges: GU 9729: calAD 860-980; GU 9730: calAD 880-990; GU 9731: calAD 880-990), Hrísheimar appears to represent a middle-to-upper status farm with an archaeofauna broadly similar in species diversity, elernent representaticn, and taphoncmy to other I cela ndic Settlement Pe riod a rc haeofa unas (McGovern et a\.2001). Since this collection is similar in terms of recovery strategy and age it provides a useful comparative baseline for the
    
    Vígishellir assembla ge. As Table 2 illustrates, Vígishellir not only has
    
    398
    
    il
    T
    
    i
    : ì
    
    Table 2.
    
    l
    l
    
    Table 2
    
    Víqishellir
    count 4683 2143 589
    6 3
    o/ /o
    
    Hrísheimar 01, context 003
    63.08 28,87 7.93 0.08 0.04
    
    sze ranoe up to 1 cm max.
    1-2 cm 2-5 cm 5-10 cm >10 cm
    
    count 233
    1
    
    o/ /ô
    
    881
    
    1501
    3BB
    
    433
    
    5.25 42.40 33.84 8.75 9.76
    
    many more very small fragments (<1 cm maximum length), but also has proportionally fewer fragments in the 2-5 cm range and almost no larger fragments. ln fact, the Vígishellir collection contains almost no whole bones -the only complete elements are carpals, tarsals, teeth, and sheep or goat toes. This seems to be the result of extremely thorough processing of the bones to elitract the smallest bits of edible marrow and collagen ("bone grease") extraction. This pattern ís characteristic of many bone collections from Norse Greenland but has not been previously observed in Viking-Early Medieval
    
    but deviates somewhat in the relative proport¡ons of element categories that comprise the LTM and MTM categories. Table 3, comparing, the propor-
    
    t¡on of long bone shaft fragments (LBF)to total LTM and MTM category counts and percentages in the Vígishellir and Hrísheimar collections, suggests that the Vígishellir archaeofauna contains a higher percentage of limb bone fragments from meat-bearing elements relative to axial skeleton elements such as vertebrae and skull fragments.
    
    lceland (0utram 1999). A substantial percentage of any highly frag. mented collection can only be identified as coming from "Medium Terrestrial Mammals" (MTM: sheepgoat-p¡g-dog size) or "Large Terrestrial Mammals" (LTM: cattle-horse size). These categories usually incorporate long bone shaftfragments (LBF), small vertebral fragments, rib fragments, small cranial fragments, and other fragments that can only be identified as mammalian and of the appropriate size class. "Unidentified mammal bones" (UNIM)are usually still more fragmented and can only be identified as mammalian rather than fish, bird, or molluscan. Higher degrees of fragmentation typically increase somewhatthe MTM and LTM proportion in a collection, while increasing the UNIM proportion even more. The Vígishellir archaeofauna follows this pattern,
    
    0vrnv¡ew oF SPECTES lo¡¡rrrrrro
    Vígishellir's archaeofauna is made up entírely of domestic mammal bones (Table 4). The MTM, LTM, and unidentified fragments were carefully searched for fish, bird, or molluscan remains but none were found. While depositional conditions in the cave were certainly atypical, they cannot account for the complete absence of these otherwise common taxa in Viking Age sites in lceland.
    
    Elr¡¡rrur DrstRrsurtoN
    The Vígishellir archaeofauna contains elements from virtually all parts of cattle, caprine, pig, and horse skeletons. Durable elements like teeth, carpals, tarsals, and the denser parts of phalanges
    
    Table 3.
    
    Tahlø
    MTM LTM
    
    A
    
    LBF 612
    159
    
    llínichøllir total
    804
    181
    
    Hrísheimar
    o/ /o
    
    LBF
    170
    7B
    
    76.12 87.85
    
    total 985 263
    
    o/ /o
    
    17.26 29.66
    
    399
    
    Table 4.
    
    Table 4
    
    Count
    
    % MSPI
    94 27 14 3 2
    
    Domestic Mammals
    (Bos taurus dom. Linn) Piq (Sus scrofa dom. Linn) Horse (Equus caballus dom. Linn) Sheep (Ovis aries dom. Linn) Goat (Capra hircus dom. Linn) Caorine (either sheep or qoat) total Caprine
    
    Cattle
    
    25.27 7.26 3.76
    0.81
    
    232
    237
    
    0.54 62.37
    63.71
    
    Total NISP
    Larqe Terrestrial Mammal Medium Terrestrial Mammal Unidentified
    
    372
    
    % TNF
    181
    
    804 6067
    
    2.44 10.83 81.72
    
    Total TNF
    
    7424
    
    and mandibles are the most common identifiable
    
    elements. lt appears that entire or nearly entire carcasses of these domestic mammals were brought to the cave, although the proporlions of long bone shaftfragments in the broader LTM and MTM categories suggest that differential transport of meatrich upper limb bones to the cave may have also taken place. lt is very clear that this is not a specialized deposit holding only low meat-value bone elements, as was the case in the slightly younger outlaw shelter, Víðgelmiç also located within the Hallmundarhraun (Ólafsson 2000), The impression given ís that the Vígishellir deposit is, if anything, skewed in favor of high meat-value elements.
    
    thin floor-like deposits within the dry stone structure and extending outward from it contained many small calcined bone fragments. lt is possible that the occupation of the structure within the cave was brief enough (or specialized enoughlthatthe sort of hearth-clearing activities regularly carried out on a normal farm did not occur here. Gnawing by carnivores (almost ceftainly dogs in lcelandl and rodents (usually mice) is regularly recorded in most North Atlantic archaeofauna. Dog gnawing has often been observed in lceland (Perdikaris et a\.20021, although it is more common in Greenlandic collections. However, the current sample from Vígishellir shows no evidence of
    
    BuRt'rtrue
    
    ,
    
    GruRwlrue, AND BUTcHERY MARKS
    
    The Vígishellir archaeofauna is also unusual in that
    
    burnt bone is virtually absent. Most lcelandic archaeofauna have a regular percentage of burnt bone that had either been deliberately used as fuel or simply thrown into the fire after meals. While the Vígishellir midden sample lacks any burnt bone, the
    
    animalgnawing. A full study of butchery practices requires a larger sample size (Lyman 1996), but qualitative observations sug gest th at the Vígishellir collection shor¡¡s unusually extreme ie'¡eis of marrow extraction and bone fragmentation. All possible sources of bone marrow were exploited - mandibles were broken open, phalanges split, even tooth roots were smashed in somc cases. Both heavy and finebladed metal tools (probably axes and knives) were used for butchery, and the many impact fractures
    
    400
    
    I
    ¡ I
    I
    I
    
    indicate the use of some sort of blunt instrument to smash open bones. The basic techniques included chopping and longitudinally splitting long bones without evidence for the distinctive dual pedoration of caprine metapodials most commonly encountered on lcelandic sites after ca AD 1100 (Bigelow 1985). ThiS level of bone fragmentation and bone processing is familiar from later medieval Greenlandic collections but is unusual in earlier lcelandic archaeofa una. Horse butchery clea rly ¡ndic ates that this species was consumed for food at Vígishellir, support¡ng a pre-Conversion (e.g. gth-1Oth century) age for the collection.
    
    pigs, cattle, and caprines, in particular, indicates that multiple individuals of different ages were consumed.
    One caprine tooth (dp4), assigned to Grant's (19821wear class "h", would conventionally be placed in the 9-12 month range (but see discussion in Enghoff 2003:54-56). lf born in May, as are most
    
    Ae
    
    r nr DrRru
    
    lcelandic sheep (Aöalsteinsson 1991), this animal most likely died in the late winter or early spring of its first year. 0ne late fetal or very newly born piglet phalanx was present, but may have been transported within a pregnant sow. Missing ent¡relyfrom the Vígishellir collection are the remains of newborn (neonatal) cattle bones. These are common on most lcelandic farm sites, reaching 40% of all identified cattle bones at Hrísheimar'
    
    The Vígishellir collection is too small and too frag-
    
    mented to allow reliable reconstruction of age profiles for any taxon. However, it is clear from the teeth preserved that fully mature cattle, pigs, horse, sheep, and goat were consumed and that some juveniles were also eaten. The tooth wear on the
    
    CoupRRRr¡vE PATTERNS
    Fig. 5 compares the Vígishellir domestic mammal
    
    collection with the other currently available lcelandic archaeofauna of the gthto early 12th centuries
    
    lceland
    
    +
    
    12tr c
    
    L z
    6 E E
    
    t¡.
    
    ñ
    
    = o
    ø
    E
    
    o
    
    ô
    
    o
    
    o
    d c õ
    
    E,
    
    i-
    
    o:
    
    E O
    'i
    
    o
    
    s
    o
    Ë :¿
    (D
    
    E
    
    Ò
    
    ;
    
    ;
    
    E O E
    
    Ê
    
    o p
    E (n
    
    c
    V
    v)
    
    c
    !-
    
    o
    
    :o
    ?
    
    Ë
    
    F
    
    u)
    
    cr
    
    I
    I
    
    D Cattle ffi Horse E Pig ØCaPrine
    Fig.5. Domestic mammal proportions in quantifiable th-lVh century lcelandic archaeofaunas
    
    401
    
    (the collections are arranged from left to r¡ght in rough chronological order). Tjarnargata 4 is from Reykjavík, Herjolfsdalur is from the Westman
    
    Ournrrs AND
    
    tNTËRPRETATIoNS
    
    lslands, Sveigakot (SVK) is a stratified collection from M'ivatnssveit, VGH (arrow) is Vígishellir. Selhagi (SLH) is a stratified collection from Mfvatnssveit, Hofstaðir (HST) is a stratified collection from Mrivatnssveit, Hrísheimar (HRH) is from M'ivatnssveit, Granastadir (GST)ís from Eyjafjörd, and Svalbarõ is from Þistílfjöt'd ín northeastern lceland (McGovern et a1.2001). The Vígishellir collection is not out of place in this series. lts mix of cattle, pigs, horses, and caprines resembles that of several middle- to upper-ranked late 9'h and 1Oth century lcelandic farms. lf we had only these data, we might conclude that this collection was from a fairly prosperous Settlement Age farm. However, as Fig. 6 shows, the Vígishellir collection is except¡onal when all of the major identified taxa are considered - no other lcelandic collection from any period consists entirely of domestic mammals and many Settlement Age collections are dominated, instead, by wild species.
    
    Surtshellir is an unusual Viking Age lcelandic site. Place names within the cave, coupled with the documented archaeological features and deposits, suggest a relatively extensive and complex suite of archaeological remains including fortifications, a subterranean house, middens, and extremely thin and fragile occupation deposits. The main habitation zone is located c. 200 meters from the cave's entrance, well beyond the penetration of any natural light. Separated from the entrance passages by a massive fortification wallin a hidden side passage, it was obviously not meant to be an accessible place. The thickness of the midden layer shows that it accumulated over some time. AMS dates confirm a late Ith or 1Oth century age for the bone-bearing deposits, while the form of the adjacent structure bears many similarities to known Viking Age halls. The massive wall spanning the main tunnel, the presence of fire-starter fragments with potent¡ally distant and diverse origins, and the unusual archaeofauna recovered from Vígishellir all
    
    +
    
    g,
    
    oqt
    
    c f (ú o
    d) (ú
    
    .c (,
    òe
    
    Pü6c>tõõõ EäJõ; ES;-+!ÞÞ rúõ*rõ'EEË tsr¡t)ll'tVi=
    
    .:Co:>ÍEE
    
    OF
    
    po CV
    >>-tnrr
    La-
    
    NN !E
    u)
    
    ØII
    
    o
    
    '¡o
    
    :f U)
    
    r1
    U)
    
    E Total Domestic El Birds E Cetacea E Seals N Caribou El Other Mammal D Fish E Mollusca
    
    Fig.6. Proportions of domestic and wild taxa in quantifiable th-lZh century lcelandic archaeofaunas.
    
    402
    
    r
    T
    ¿
    
    match reasonably well with correlates one might reasonably derive from medieval accounts of outlaw bands' lairs. The 2001 archaeofauna from Vígishellir, while small and fragmented, shows both similarities and differences from other currently known lcelandic Settlement Age and Early Commonwealth samples. It would'be anomalous (except in relative proportions of domestic mammals) even if it did not come from a cave with such a colorful legend. lt lacks any of the fish, birds, and shellfish common on Settlement Age sites in lceland, suggesting that the cave's occupants were not reduced to scavenging wild resources despite the site's marginal location - 10-15 meters underground in a lava field on the fringes of lceland's uninhabitable interior highland. Recent work indicates that farms 60 km or more
    
    they lacked otherfood sources (dairy produce, fish, birds, cereals), or possibly they attempted to limit their exposure to community retaliation by spacing their raids as widely as possible. Given the density of this midden and its composition, it would appear that the occupants of the cave must have had a heavy impact on the economies of the farms around
    them.
    Yet, while Vígishellir's occupants built significant structures and dined on the meatiest portions of
    
    from the coast were regularly provisioned with preserved marine fish, sea birds, sea mammals, raw materials, and other critical resources from the Settlement Period onwards (Smith 2004; Amundsen ef a/. in press). lf the cave's occupants were indeed outlaws, they may have been cut off from regular access to some resources because they no longer had access to the social networks that allowed such provisioning. The mix of domestic mammals present suggests that the cave's occupants had the ability to acquire a wide range of domestic stock from surrounding farms, due perhaps to their success as raiders or to support, willing or coerced, from nearby farms. The absence of newborn calf bones (so common in farm middens) may pointto seasonality in their raiding activity, problems in capturing young animals normally kept withiñ farmyards, or a simple focus on adult animals that could provide more meat when slaughtered' The apparent surplus of meat-rich long bones could reflect the butchery of some captured animals away from the cave, or raids on farm smokehouses or meat stores. And yet the pattern of bone f ragmentation also suggests that while the cave's occupants may have
    
    many animals, crushing their bones while burning few and carefully separating the burned from the unburned, these patterns differ significantly from those recorded at Viðgelmir, another suspected 1Oth century outlaw shelter in the Hallmundarhraun (Ólafsson 2000). At Viðgelmir, an untidy scatter of burned and unburned bones, tossed unsorted around a small unframed hearth on a hidden ledge testifies to a short stay, perhaps lasting no more than a few days, by one or two people eating the least desirable portions of a cow (Ólafsson 2000)'
    That these two nearly contemporaneous sites, located no more than t0 kilometers from one another are so different, suggests, at the very least, that the archaeological records of outlaws and
    
    cave sites in lceland may be more complex than we
    
    currently realize.
    
    AcTTowLTDGEMENTS Fieldwork in Vígishellir, radiocarbon dating and trace element analyses were undertaken with supportfrom the National Geographic Society's Committee for Research and Exploration, Þjóðminjasafn íslands, the Buffalo Museum of Science, and the Haffenreffer Museum of Anthropology (Brown University). Archaeofaunal analyses undertaken at N0RSEC (CUNY-Hunter), reported here, are the product of sustained supportfrom the UK Leverhulme Trust "Landscapes Circum Landnám" proiect, the US National Science Foundation, the lcelandic Science Council, National Geographic Society, PSCf\r !Ntv Granrc Prnnrarn. and Þióðminiasafn íslands qrrq rrvv¡¡r¡¡r ¡ lvYrurr¡r ulsl¡lr uulll (the lcelandic National Museum). The authors jointly wish to express their gratitude to all of the above spollsors Toi pasl anu pf cseili suplrut r'
    
    enjoyed considerable success in carrying off ^-i-^t domestic stocit, titey ilad tû proüess an¡mal carcasses completely for meat, marrow, and bone grease. Perhaps they were attempting to get the rnost out of the meat animals they car-rgi'rt hecause
    
    403
    
    H
    R
    ¿
    
    REFERENCES
    Aðalsteinsson,
    
    ¿
    
    ù
    s
    '¿,
    
    S.
    ín early lcelandic
    1
    
    Jónsson, G.
    
    i
    n ga
    
    1991 The importance of sheep
    
    agriculture.
    
    1953 Harðar saga ok Hólmverja. fslendinga sögur,vol.12.
    ísle ndi
    sa
    g na
    
    Acta Archaeologic a 61:285-29
    
    .
    Krivogorskaya,
    
    útg áfa n, Beykiavík.
    
    Amundsen_. C. P., Perdikaris, S., McGovern, T. H.,
    Y..
    
    Jónsson, S. S. and Hróarsson, B.
    
    Brown, M., Smiarowski, K., Storm, S., Modugno,
    
    S.,
    
    1992 Fornminjar ííslenskum hraunhellum. Surtu7 ársrit Hellarannsóknafélags lslands 1992:10-2?.
    
    Frik, M. and Koczela,
    
    M. NorthArchaeology. ísfirðinga,
    1996
    
    ln press Fishing booths and fishing strategies in medieval lceland: an archaeofauna from the site ofAkurvík, Laxness, H. K.
    
    West lceland. Environmental Árnason,
    
    1949 Lítil samantekt um útilegumenn.
    ninga r, Maî 1 949: 86-1 30.
    
    Tímarit Máls og men-
    
    B.
    
    1971 Yfirskyggðir staðir. Ímsar athuganir. Aldur hellismanna:
    96-105. Reyklavík.
    
    1889
    
    Saga Hellismanna, fyrsta prentun, Prentsmiðju
    
    Ísafjörður.
    Lyman, R.L.
    
    Bigelow G. F.
    
    Taphonomy. Cambridge University Press, Cambridge.
    
    1985 Sandwick,
    
    Unst, and the Late Norse Shetlandic economy.
    
    ln Smith, B.led.l Shetland Archaeology, New Work
    
    in 95-
    
    McGovern, T.H.
    
    Shetland in the 1970b. Shetland Times Press, Lerwick:
    
    1992 Bones, buildings, and boundaries: paleoeconomic
    approaches to Norse Greenland. ln Morris, C.D. and J. Rackham (eds.) Norse & later Settlement &. Subsistence
    
    "127.
    Enghoff, l.
    
    B. Fãrm
    2001
    
    ¡n the
    
    Nofth Atlanfic. Glasgow University Press, Glasgow;
    
    2003
    
    Hunting, fishing, and animal husbandry atthe
    
    157-186.
    
    Beneath the Sand, Western Greenland: an archaeozoo-
    
    Seülemenl. Meddelelser om Grønland, Man & Society 28, Copenhagen.
    logical analysis of a Norse farm in the Western
    Gestsson, G.
    
    McGovern T. H., Perdìkaris, S. and ïnsley, C. Economy of Landnám: the evidence ofzooarchaeology. ln Wawn, A & Sigurðardóttir, Þ. (eds.) App roaches to Vínland. Sigurður Nordal lnstitute Studies 4, Reyklavík: 154-166.
    
    i960
    
    Hallmundarhellir. Árbók hins íslenzka
    1960:
    
    Fornleifafélags
    
    McGovern, T.H. and Perdikaris, S.
    
    76-82.
    
    2002 Preliminary Beport
    
    of Animal Bones from Hrísheimar, N.
    
    lceland, report on file with Fornleifastofnun l'slands and
    Grant,
    
    A.
    a
    
    National Museum of lceland, BeykiavÍk,
    
    1982 The use oftooth wearas
    .
    
    guidetothe age of domestic
    Ólafsson,
    E.
    
    ungulates. ln Wilson, 8., Grigson, C. and Payne, S. (eds.) Ageing and Sexing Animal Bones from
    Srres. BAR British Series 109, 0xford:
    
    Archaeological 91-t08.
    
    1978 Ferðabók
    
    Eggerts 1lafssonar og Bjarna Pálssonar.Orn
    
    og örlygur, Reykjavík.
    
    Grönvold, K., Óskarsson, N., Johnsen, S. J., Clausen, H., mer. C. U., Bond, G. and Bard,
    
    Ham-
    
    Ólafsson, G.
    
    E. 149-155.
    
    2000
    
    Fylgsnið í hellinum Víðgelmi. Árbók hins íslenzka Forn-
    
    1995 AshlayerfromlcelandintheGreenlandicGRIPicecore leifafélags1998:125-142.
    correlated with oceanic and land sed¡ments. Earth and Planetary Science Letters 135:
    Hróarsson, Ólafsson, G., Smith, K.
    P.
    
    and Stefánsdóttir, A.
    
    20û4 frannsókn
    
    á miniúm í Surtsheiii. Bannsóknaskyrslur Þjoð-
    
    B.
    á ístandi. Reykjavik.
    
    minjasafns1}}l, Vlll. Þjóðminiasafn íslands, Reykjavík.
    0utram, A.
    K.
    
    1990 Hraunhellar
    Jóhannesson,
    
    H. Nát9,
    
    1999 A comparison of Palaeoeskimo and medieval Norse bone
    fat exploitation in Western Greenland. Arctic Anthropology,36l1l:103-l
    17.
    
    19Bg Aldur Hallmundarhrauns íBorgarfirdi. Fjölrit
    túrufrædistofnunar
    
    404
    
    ,w
    ¿
    
    I
    
    Pálsson, H. and Edwards,
    
    P.
    
    (trans.)
    
    Smith. K.
    
    P.
    
    1972
    
    The.Book
    
    of Setttements (Landnámabók)' University of
    
    2004
    
    Manitoba lcelandic Studies, Vol' 1' University of Manitoba Press, WinniPeg. Perdikaris, S., Amundsen, C. and McGovern,
    
    lndependent people, householders and outlaws: reconciling economic autonomy, political centralization, and trade
    in medieval lceland. Paper presented atthe 69th Annual
    
    Meeting of the Societyfor American Archaeology, Montréal'
    
    I
    
    H.
    
    2002 Reportof Animal BonesfromTjarnargata3C,
    
    Reykjavík, lce-
    
    Þórðarson, M.
    
    land, report on file with Fornleifastofnun íslands, Reykiavfk'
    
    1910 Tveir hellar í Hallmundarhraunl
    
    Skírnir' 331-351'
    
    405
    
    Dynamics of Northern Societies
    Proceedings of the SILA/NAB0 Conference on Arctic and Nofth Atlantic Archaeology, Copenha gen, May 1 0th-1 4th, 2004
    
    Edited by Jette Arneborg & Bjarne Grønnow
    
    SILA
    
    -The Greenland Research Centre
    
    at the National Museum of Denmark
    
    s-*ffi
    
    PNM
    Publications from the National Museum Studies in Archaeology & History Vol. 10 Copenhagen 20û6
    
    Publications from the National Museum Dynamics of Northern Societies Proceedings of the SlLA,/NAB0 Conference on Arctic and North Atlantic Archaeology, Copenhagen, May
    1
    
    Oth-14ú, 2004
    
    Editors ofthe PNM series: Poul Grinder-Hansen & Lars Jørgensen Copyright The National Museum of Denmark and all individual authors All rights ieserved Edited by Jette Arneborg and Bjarne Grønnow Technical edition: Niels Alqreen Møller
    
    Linguistic revision: Aoife Daly Cover design by Marianne Blank Produced by Special-Trykkeriet. Byens Tryk a-s Published by The National Museum of Denmark 12, Frederiksholms Kanal. DK-1220 Copenhagen
    ISBN: 87-7602-052-5
    
    Edito.rial advisor::.lrlqll" Appelt colleen Barey, H.c. Gulløv, Einar Lund Jensen, Ditlev Mahler, Jom Mc{jovern, Ulla odgaard, Anne Pedersen, Mikkel Sørensen. 0rri Vesteinsson
    
    K
    
    ISSN:0909 EAN code: 97
    
    8877 6020521
    
    This book is available direct from: The National Museum The Museum Shop 12, Frederiksholms Kanal DK-1220 Copenhagen K
    Fax: + 45 33473330
    
    shop@natmus.dk
    
    The Proceedings are funded by: SILA - The Greenland Research Centre ât the National Museum of Denmark The Department of Research and Exhibition, National Museum of Denmark
    Cover: Hatherton Bugt, lnglefield Land, 1997.Original photo by Bjarne Grønnow

Add Comment

To add a comment, you need to login or signup (you can signup without leaving this page)