Note: This paper was written for the proceedings of the conference ‘Fishery, Trade and Piracy. Baltic and North Sea in the Middle Ages and later’, 11-13 May 2006, Tammisaari (Finland). As publication seems very unlikely at this point (2012), I am offering this paper online. An island archaeological approach to the Viking colonization of the North Atlantic Pieterjan Deckers Introduction The present paper is a brief exploration of the application of methods commonly used in the archaeological study of the Pacific and Mediterranean islands to the expansion of the Vikings across the North Atlantic during the ninth to eleventh centuries AD. The pillars of this approach will be island biogeography, combined with a theoretical understanding of the processes of migration and island colonization. W.F. Keegan and J.D. Diamond (1987) were the first to realize the full potential of a biogeographical methodology in the archaeological and historical study of islands, although some island archaeologists were already using rudimentary versions of Keegan and Diamond’s methods before that date (notably Cherry 1981, 1985). Since 1987, this approach has been applied to numerous episodes of island colonization, mainly in the Pacific (e.g. Bowdler 1995; Irwin 1996) and Mediterranean (e.g. Patton 1996; Bass 1998; Broodbank 1999, 2000). It has to be emphasized that this paper is intended as an intellectual experiment rather than a definitive statement about the course of the westward expansion of the Vikings. It originates from a licentiate thesis at the Catholic University of Leuven (Deckers 2005) consisting primarily of a critical evaluation of the theories and assumptions current in the archaeological study of island societies. The case-study on which this paper is based was intended to test these theories in a particular and well-documented sequence of island colonizations. Due to these limitations and because of the size of the study area, it was not possible to base this paper on a complete and up-to-date status quaestionis of the archaeological, historical and palaeoecological research in the North Atlantic. Although thus restricted in its scope, this paper aims to demonstrate the possibilities biogeographical methods offer in drawing together the evidence from as large a geographical expanse as the Norse ‘colonies’ in the North Atlantic into one coherent picture. Before moving on to this analysis, the extant literary, archaeological and other evidence for the Viking exploration and colonization of these places have to be briefly reviewed. Viking explorers and settlers in the North Atlantic The earliest documented westward ventures of the Vikings outside their Scandinavian homeland are the raids on several religious communities in Scotland and England in the last decade of the eighth century AD. From that time on, Orkney, Shetland, the Faroes, Iceland and Greenland were rapidly settled by Norse migrants. Within two centuries, the Norse had reached the North American continent and founded a (short-lived) settlement on Newfoundland. Orkney Orkney, the southernmost archipelago within the area of study, is separated from the Scottish mainland by the Pentland Firth, a strait of barely 10 km wide. The soil is relatively fertile, and of all the Norse colonies in the North Atlantic, agriculture had by far the greatest importance in Orkney. Except for sandstone suited for building, natural resources are very scarce (Morris 1985: 215; Omand 1993: 102-104). -1- 8/01/2007 h:01 a1/p1 It is not very likely that the Vikings who raided the Scottish coasts from the last decade of the eighth century did so without passing by Orkney. Finds of steatite and reindeer antler in Orkney (Barrett 2003: 80) and of Insular objects in Norwegian burials (Myhre 1993: 191; 1998: 8) suggest that there was already some contact and trade between the British Isles and Scandinavia in the decades (centuries?) before the first raiding expeditions. The islands themselves may not have had a lot to offer to Viking raiders, but they are located in a strategic position controlling the access to the rich monasteries and trade ports in the south. It can be reasonably assumed that temporary military encampments appeared in Orkney, as elsewhere in Europe, from the first half of the ninth century. These encampments may have transformed into more permanent settlements in the course of that century (Crawford 1987: 40; Owen 2004: 22-23). The archaeological finds of settlements confirm a ninth-century date for the earliest stages of Norse occupation, but also show how the Viking immigration in Orkney is a complex phenomenon. Material remains of Norse and indigenous Pictish origin are found together at almost every early Viking settlement, and this makes establishing the date of the earliest arrival of the Scandinavians a precarious task. In Pool (Sanday) the Pictish-Norse transition has been dated to the end of the eighth or the beginning of the ninth century (Hunter et al. 1993: 275, 280-281). Other Viking settlements in Orkney most probably date from the second half of the ninth or the tenth century (Graham-Campbell & Batey 2001: 160-171). Viking burials (with grave goods of Scandinavian origin) do not appear before 850 AD, providing a terminus ante quem for the arrival of the Vikings (Graham-Campbell & Batey 2001: 57-58). The co-occurrence of indigenous and Viking cultural elements in the earlier stages of colonization is in stark contrast with the eventual outcome of the colonization: the complete domination of Norse toponymy, language and culture by the end of the ninth century. Most authors now consider a gradual process of submission and assimilation to be the most likely course of events (e.g. Barrett 2003). Scandinavian merchants, warriors and farmers started settling permanently in the first half of the ninth century and seized power from the Pictish aristocracy. From 850 AD on, Orkney occupied an increasingly important position in the Norse trade network extending across the British Islands and the North Sea and became an important political power. Later literary sources (e.g. Orkneyinga Saga), although not objective historical reports, give the impression of a reorganisation and consolidation of existing Norse settlement in the second half of the ninth century AD (Morris 1985: 211-213), perhaps because of a larger influx of Norse migrants (Owen 2004: 23-32). Shetland Shetland is situated closest to Scandinavia of all island groups in the North Atlantic. Typical for the landscape are the lack of trees and the widespread layers of peat of up to six meters thick. The landscape is extremely moist and best suited for extensive pastoralism (Turner 1998: 13- 15). Not much is known about the early Viking period in Shetland, but there are many similarities with Orkney. Here too, several settlements show a transition period in which Pictish and Viking elements appear together (Stummann Hansen 2000: 98; Owen 2002: 160), for instance at Old Scatness (Turner et al. 2005). Unfortunately, this transition phase has only been roughly dated at the earliest in the ninth century (Graham-Campbell & Batey 2001: 64-67). The oldest Norse objects from graves were typologically dated to the ninth and tenth centuries AD (Shetelig 1954: 69; Stummann Hansen 2000: 89). The treasure from St. Ninian’s Isle, a collection of silver objects of Pictish origin, was accurately dated to around 800 AD. The suggestion that the deposition of this treasure was a precaution against Viking raids remains highly hypothetical, but the presence of wealthy Picts in Shetland may have been enough to draw Viking attention (Turner 1998: 94; Graham- Campbell 2001: 69). -2- 8/01/2007 h:01 a1/p1 According to Graham-Campbell and Batey (2001: 156), the archaeological remains do not indicate Scandinavian settlement in Shetland before the middle of the ninth century. Nevertheless, given its location intermediate location between Scandinavia and the British Isles, this does not preclude earlier contacts or (temporary) settlement going back to at least the beginning of the ninth century. The Faroe Islands The Faroe Islands are situated approximately halfway between Shetland and Iceland. The relief is varied and more rugged than Orkney or Shetland. Trees are absent and the landscape is dominated by low heather. Today as in Norse times, sheep are the Faroes’ most important economic asset, and they possibly gave the archipelago its name. Agriculture is limited to the narrow coastal strip and the valley bottoms (Jackson 1991: 17-18; Edwards 2005: 586-587). If Alan Small’s (1969: 149) model of settlement requirements is applied to the Faroes, only 7% of the total area was suited for settlement in Norse times (MacGregor 1984: 3). Dicuil, an Irish monk working at the Carolingian court in Aachen, wrote around 825 AD that Irish religious men used to live on a group of islands usually identified as the Faroes. By the time of writing however, this settlement had been destroyed by what he calls ‘Norse pirates’ (Debes 1993: 458-459; Stummann Hansen 2003: 56). The results of recent palaeoecological research further suggest human occupation from the end of the 6th century, but this has yet to be confirmed by archaeological evidence (Arge et al. 2005: 599; Dugmore et al. 2005: 25; Edwards 2005: 588). If there was indeed an Irish presence on the Faroes, information about the islands may have been transferred to Scandinavians in Scotland as early as the seventh or eighth century. However, these few Faroese hermits could hardly have been an attractive target for plundering Vikings. Consequently, the Scandinavians sailing to the Faroes were looking for land, not for riches (Stummann Hansen 2003: 60). The earliest mention of the Faroes in Scandinavian written sources can be found in the Faereyinga Saga. The reliability of the saga is questionable (Halldórsson 1961: 51; Foote 1970: 159-160), and the chronology of early events is unclear. Nonetheless, the following hypothetical scenario can be constructed. The first Norse settler arrived in the Faroes in the first half of the ninth century (his grandson took part in the landnám of Iceland in 874). Towards the end of the ninth century, when many opponents of Harald Fairhair fled to the islands, the colonization accelerated (MacGregor 1984: 11; Debes 1993: 457-458; Stummann Hansen 2003: 33, 44). As is the case with Shetland, the archaeological knowledge of the early Viking period in the Faroes is rather limited. Finds from the earliest Viking burials, such as those at Yviri í Trøð and Við Kirkjugarð (Sandur, Sandoy), are typologically dated no earlier than the tenth century (Dahl 1965: 137; Arge & Hartmann 1992: 5; Arge 1993: 471; Stummann Hansen 2003: 45; Arge et al. 2005: 612). The earliest known carbon date from the Faroes is the interval 760-890 AD from hazelnut shell in a floor layer at the farm site of Toftanes (Leirvík, Eysturoy) (Edwards 2005: 591). Similarly early carbon dates come from the sites of úti á Bø at Sandur, Sandoy (Arge et al. 2005: 612- 613) and Argisbrekka on Eysturoy (Mahler 1991: 67-72; 1993: 494). The latter, a shieling or seasonal herding settlement, was in use from the first half of the ninth century to the middle of the eleventh century. It can be concluded that the Faroes were settled by the middle of the ninth century, even if only by relatively small numbers of immigrants. Iceland The inland area of Iceland consists mainly of a plateau unsuitable for human occupation. Only the coastal areas and the large coastal plains of the southwest were suitable for Norse -3- 8/01/2007 h:01 a1/p1 occupation. Agriculture is restricted to only a minimal fraction of the total surface area (Buckland et al. 1991: 259; Smith 1995: 323). Dicuil and other written sources mention the presence of Irish clerics on Iceland. This would have important implications for our understanding of the process of Norse expansion across the North Atlantic, but, as on the Faroes, no firm archaeological evidence for a pre-Norse presence has been found (Ahronson 2000, 2002, 2003; Sveinbjarnardóttir 2002). Two written sources from medieval Icelandic literature, the Islendingabók and the Landnámabók, contain a history of the settlement and the families of the first settlers in relation to the landscape. Notwithstanding the difficulties of distinguishing historical fact and fiction in these works, the Islendingabók in particular is considered an invaluable source of information on the Norse colonization of Iceland (Friðriksson en Vésteinsson 2003: 145-146). The Islendingabók and the Landnámabók recount how Iceland was coincidentally discovered around 860 AD, how a first colonist tried in vain to settle in 865 AD, and how a better- organized and large-scale colonization from Norway started in 870 or 874 AD. Around 930 AD, when the Althing was established, the whole of Iceland was occupied according to these sources (Smith 1995: 320; Friðriksson en Vésteinsson 2003: 142). Notwithstanding some scepticism towards the literary sources and archaeology’s reliance on this category of evidence (e.g. Friðriksson en Vésteinsson 2003: 157-158), the present consensus remains the traditional view of a large-scale colonization starting in the last decades of the ninth century. Archaeological remains are generally confined to the strata above the landnám tephra dated at 871 ± 2 AD (Grönvold et al. 1995), and most of the inhabitable parts of Iceland were indeed occupied by 930 (Vésteinsson 2000: 167-168; Stummann Hansen 2003: 59-60; Dugmore et al. 2005: 25). It is estimated that between 8.000 and 20.000 settlers had arrived by that time (Steffensen 1975: 446; Helgason et al. 2001: 733). Greenland Despite the relatively high temperatures of the Medieval Optimum, only the warmer southwestern part of Greenland’s ice-free coastal strip was suitable for the pastoral economy of the Viking settlers. Agriculture was impossible at all times in Greenland, but this shortage was compensated by the rich hunting and fishing grounds (McGovern 1980: 249-250; Lynnerup 2000: 290). Even so, the Norse colonies in Greenland were never self-sufficient and relied on trade with Europe throughout their existence (Arneborg 2003: 171). The colonization of Greenland is recounted in two very informative written sources, Grœnlendinga Saga and Eirik’s Saga, complemented by data from other sources, such as the already mentioned Landnámabók and Islendingabók (Magnusson & Pálsson 1965: 29-31; Arneborg 2003: 165, 166). Greenland is said to have been discovered by coincidence around 920 AD, and Icelanders landed on Greenland for the first time in 981-982 AD. With Eirik Thorvaldsson (better known as Eric the Red), the traditional history of Greenland took a decisive turn. In 982 AD Eirik was accused of murder and banished from Iceland for three years, during which time he thoroughly explored the new land. Upon his return, he recruited settlers to found a new society in what he somewhat euphemistically called Grönaland. Eirik himself, along with ten other high-ranking settlers and their followers, settled in the so-called Eastern Settlement (at the southernmost tip of the west coast), while other colonists founded the Western Settlement 400 km further to the north (Magnusson & Pálsson 1965). In total, the initial group of settlers is estimated to have numbered four to five hundred individuals (Lynnerup 1996: 123). This colonization took place at a time when the autochthonous Dorset- and Thule-Eskimos were living further north in Greenland and the Canadian Arctic. Only from 1200 AD onwards, at the beginning of the Little Ice Age, Thule groups gradually moved southwards and increasingly -4- 8/01/2007 h:01 a1/p1 came into contact with the Norse (Saillard et al. 2000: 719, 724; Gulløv 2000: 323-324; Schledermann and McCullough 2003: 198-199). The archaeological knowledge of this colonization process is fragmented, but it corroborates the literary evidence. Only a few sites can be dated to the earliest stages of Norse occupation with certainty (Keller 1991: 130-131; Lynnerup 2000: 290-292). One of the most important is the recently excavated farm of Gård Under Sandet (GUS) in the Western Settlement. The oldest phase, a turf-walled longhouse, has been dated to the first half of the 11 th century both by AMS- dating and architectural typology (Berglund 2000: 295-298; Albrethsen 2003: 107). Another early site is Thjodhild’s Church, a chapel and graveyard near Eirik’s farm at Brattahlid. According to Eirik’s Saga, this is the first place of Christian worship to be built in Greenland after the introduction of Christianity around 1000 AD. Carbon dating of some of the buried bones produced a slightly older date at the end of the tenth century, which might indicate that the church was already in use before 1000 (Arneborg et al. 1999; Arneborg 2000: 310-311). Vínland The study of Viking presence in Vínland differs in several aspects from the other regions covered in this paper. Firstly, the main sources of evidence (two very similar texts and one archaeological site) are very limited in number. However, the information these sources yield is much more complete than the information available for any other aspect of early Viking activity in the North Atlantic. Secondly, Vínland was not actually colonised, and the Viking activities in this region mainly consisted of exploration and exploitation of the north-eastern coasts of North America. In this respect, the case study of Norse Vínland is a crucial one: the detailed information obtained here concerning the stages preceding actual migration and permanent settlement might help to fill some of the gaps in the knowledge of the colonization process of other regions in the North Atlantic. The principal written sources are the already mentioned Grœnlendinga Saga and Eirik’s Saga, often referred to as the Vínland Sagas. They tell the story of the coincidental discovery of forested coasts to the southwest of Greenland by a merchant on his way to Greenland in 986 or 987. Leif Eiriksson decided to mount an expedition to explore these lands, which must have been very attractive from the viewpoint of treeless and cold Greenland. Leif successively visited Helluland (‘Land of the Flat Stone’ or Baffin Island and the north of Labrador), Markland (‘Woodland’ or central Labrador) and Vínland (‘Wineland’, where the Norse discovered vines). After a winter spent at the newly founded Straumfjord in Vínland, the expedition returned home with a shipload of lumber, hides and grapes. Several further expeditions were organised to exploit the natural riches of Vínland during the summer, and Straumfjord remained in use as a winter camp. One of these expeditions had the explicit intention to found a permanent settlement. However, the Norse repeatedly came into conflict with indigenous Native American tribes. Combined with the long distance from Greenland and the drain of manpower these expeditions meant to the small settler population of Greenland, this put an end to the Norse expeditions to Vínland not even two decades after the first exploration (Magnusson & Pálsson 1965; Seaver 2000: 270-271; Wallace 2000: 213). In the 1960’s and ‘70’s a Viking settlement was excavated at L’Anse aux Meadows at the northernmost tip of Newfoundland. This settlement acted as a gateway to the region around the Gulf of St. Lawrence, probably the location of Vínland proper (Wallace 1991: 194; Wallace 2000: 213-216). The archaeological site of L’Anse aux Meadows consists of eight buildings grouped into three complexes, each comprising a large hall with rooms to the sides and a couple of secondary buildings (Wallace 2000: 209; 2003a: 225-226). It is likely that every hall housed one ship’s crew. The settlement is carbon dated around 1000 AD with a high level of certainty (Wallace 1991: 182-183; 2003a: 226; 2003b: 168-169). Several arguments, such as the absence of a graveyard, the presence of only one construction phase and the relatively limited amounts -5- 8/01/2007 h:01 a1/p1 of waste, bear witness to a very short occupation period. Almost no utensils or other objects were recovered, pointing to the organized and orderly way in which the settlement was abandoned. All this makes sense in light of the written sources (Wallace 2000: 213-214; 2003b: 169-170). Thus, according to Wallace (2003a: 231-232), we can confidently identify L’Anse aux Meadows as Straumfjord. The larger picture: island archaeology in the North Atlantic The available archaeological, literary and other evidence from the individual islands and island groups seem to convey a relatively complete image of the Viking expansion in the North Atlantic. However, an insight in the motivation of the migrants and a grand perspective of the expansion in its whole are still lacking. For that purpose, not only the circumstances in the homeland during and immediately prior to the emigration need to be taken into consideration, but the islands themselves – their geography and archaeology – need to be analysed more profoundly as well. This analysis will result in an assessment and mutual comparison of the attractiveness, chronology and course of the colonization process of the different islands. A better understanding of the Viking expansion in the North Atlantic, exceeding the conclusions that can be drawn from individual case studies, will be the final outcome. The basic structure of migration Only since the last two decades or so, migration has been rehabilitated as a valid way of explaining the archaeological record (Anthony 1990; Chapman & Hamerow 1997; Burmeister 2000). As a result of anthropological, sociological and geographical research on migration, it is now a very well understood phenomenon (e.g. Agozino 2000; Bretell & Hollifield 2000). The push-pull model (Lee 1966, Dorigo & Toller 1983) is one of the classic paradigms for the analysis of migrations. Although it has its critics (e.g. Amin 1995: 31-32; Malmberg 1997: 29- 30), it is still useful as a heuristic tool. According to this model, any migration consists of two main elements: a place of origin and a destination. Both origin and destination have certain characteristics, which can be positive (so- called pull-factors) or negative (push-factors) (fig. 1). The decision whether to migrate or not, is mainly determined by weighing the positive aspects of origin and destination against the negative. Migration is thus stimulated by pull factors in the place of destination and push factors in the country of origin (Lee 1966: 50). The decision to migrate is also influenced by the migration route, the obstacles one has to overcome to reach one’s destination. Geographical distance is undoubtedly the most important of these obstacles (Lee 1966: 49). If looking to answer the question ‘what drove the Vikings to settle on the seemingly barren islands of the North Atlantic?’ it is thus essential to examine potential push and pull factors as well as the migration route. Push factors in the homeland? As a first step in the analysis of the Norse migration, the push factors in the homeland should be considered. Recent studies, mainly based on genetic evidence (e.g. Helgason et al. 2000a, 2000b, 2001; Wilson et al. 2001; Jorgensen et al. 2004; Goodacre et al. 2005), have tended to emphasize the non-Scandinavian, British origin of part of the settlers in the Faroes and Iceland. Nonetheless, Norway is still generally regarded as the primary country of origin of the North Atlantic settlers. Some circumstantial factors were indirectly influential, such as advances in seafaring technology (Christensen 2000: 87-88; McGrail 2001: 210-212) and the improving climate at the start of the Medieval Optimum (Schneider & Londer 1984). Although these factors undoubtedly stimulated seafaring and improved the chance of successful settlement, they can hardly be regarded as more than facilitating. -6- 8/01/2007 h:01 a1/p1 Traditionally1, a population increase is said to be the main cause of emigration to the seemingly inhospitable islands of the North Atlantic. In particular, the fertile southwest of Norway was quite densely settled by the end of the early Iron Age (Helle 1998: 250). However, there are few archaeological indicators of economic intensification or ecological stress. On the contrary, marginal grounds in Norway are only brought into cultivation from the tenth century on, well after the heyday of the expansion (Myhre 1998: 11-17; 2000: 36-37). The beginning of the expansion coincided with the rise of the Carolingian empire and the peak of Irish and British missionary activity. A conflict for control over North Sea trade arose, but the ideological-religious differences between the Christian kingdoms of Western Europe and Scandinavian-Germanic paganism appear to have had an influence as well. Although opportunism was without doubt an important factor, the raids by Norse ‘pirates’ cannot exclusively be attributed to individual aspirations of wealth and fame (Myhre 1993: 195-198; Myhre 2000: 44; Näsman 2000: 5-6). Things were changing in the Viking homelands as well. At least from the ninth century onwards, Norway was marked by a growing political centralisation. The traditional view is that Harald Fairhair was king of a united Norway after his victory at Hafrsfjord in 872 AD. Although there is probably some truth to this story (Helle 1998: 253-254; Sawyer 2003: 32), in reality the process of centralisation was not completed until almost two centuries later (Roesdahl 2003: 158). In the same period, Christianity is introduced in Scandinavia as an element of the political strategy of the ruling class (Roesdahl 1992: 36). Perhaps the most profound and disrupting change happened in the socio-economic sphere. Until the ninth century, Scandinavian economy was based on prestige. The purpose of valuable objects was to generate and signal social status in a lively system of gift exchange. War and plundering were ideal ways for chiefs to fulfil the need of movable wealth to redistribute among their warrior retinues. Along with the establishment of centralised authority, however, the first trade ports in Scandinavia were founded, coinage appeared, and Scandinavia was integrated in the Western European trade networks. Valuable objects lost their social significance and the prestige-economy was replaced by a market economy (Lagerqvist 1992: 220-221; Hedeager 2000: 85-86). Other contributing factors have been proposed, however. Emigration may have been stimulated by the growth of a poor class of free farmers and the particular hereditary laws of the Vikings (Skre 2001a: 12 – but this view is challenged by Lunden et al. 2001: 14, 18). In fact, the tribal structure of Viking society in its whole, with its large aristocracy, strong concept of honour and great social mobility, probably favoured emigration as a solution to social conflict (Skre 2001b: 26). This may explain why the emergence of centralised states triggered such a large-scale emigration. The Orkneyinga Saga (Pálsson & Edwards 1981: 26) relates how a considerable number of Harald’s adversaries fled to the islands, whence they raided the Norwegian coasts during summers. For the defeated chiefs, emigration was a way to retain their former power and status. Some (e.g. Keller 1991: 128) even suggest that the expansion to the west was a deliberate attempt from a conservative faction of Norse society to save the traditional way of life of the Vikings at a time when it was disappearing in the homeland. However, the structural approach of the above paragraphs may reveal only part of the story. Every individual migrant made the decision to migrate on basis of a personal consideration of circumstances. For instance, it is a well-known tradition that Eirik the Red was an exiled murderer, but he is certainly not the only case mentioned in written sources. Ingolf Arnarson, one of the first settlers of Iceland, fled from Norway on accusation of murder (Thorláksson 2000: 177). 1 In fact, this tradition can be traced back to at least the 6th century AD, when Scandinavia was described as ‘vagina nationum’ and ‘officina gentium’, the birthplace of peoples, in Jordanus’ Getica (Boyer 1992: 77). -7- 8/01/2007 h:01 a1/p1 Apart from political, economic, social and personal factors, other elements potentially played some part in the colonization of the North Atlantic. From the already cited observation that marginal lands in Norway were only taken in use towards the end of the expansion, it could be inferred that the Viking colonists preferred remote and barren islands above less fertile parts of their own homeland. The historical events in Norway explain this at least partly, but psychological factors may have played a role as well. Overseas travels were an important source of social prestige for members of the Viking aristocracy, not only through the profits of trade and looting, but because of the considerable investment the equipment and crew of a ship represented. The control over sea routes along the Norwegian coast and across the North Sea was of great political importance. It is possible that a mentality of orientation towards overseas regions originated from this familiarity with and dependence upon the sea. This might explain that islands are colonised, even if these islands in our eyes do not seem to be advantageous when compared to marginal grounds on the mainland. From the anthropological study of migration, it is known that previous experience in travel and migration lowers the threshold for further migration (Lee 1966: 54). The experience of migrating, a sense of ‘unboundedness’ or ‘out-of-placeness’ (Bauman 1995: 95) and the consciousness of origins elsewhere may become part of the social identity of a migrant and form a stimulus for further migration (e.g. Jones 1999). This brings into recollection the exploration and colonization ideology proposed by Keegan and Diamond (1987: 67-69) and Broodbank (1999: 28) to explain parts of the colonization process of the Pacific and the Aegean islands respectively. According to these authors, the colonization of these islands was at least partly driven by the success of previous colonizations and by a common history involving overseas migration. The above-mentioned phenomena all have to be taken into account as part of the context of the migration of Norwegian Vikings to the islands of the North Atlantic in the ninth and tenth centuries. A causal relation cannot be pre-assumed, but it is reasonable to suppose that at least some of these factors have influenced the colonization process. Pull factors in the islands? A biogeographical analysis of island attractiveness The second issue requiring analysis is what pull factors existed in the destination of the Viking migration, the islands of the North Atlantic. At first sight, the landscape and climate of these islands are very similar to Norway. However, when compared to the economic resources of Scandinavia, the islands definitely constitute a poorer environment: many were ill suited for agriculture, virtually treeless and lacking in important mineral resources. Nonetheless, apparently Norse settlers found them attractive enough for colonization. Quantitative biogeographical methods enable an assessment of the attractiveness of individual islands and archipelagos at a more detailed level. As will be demonstrated further on, these methods might help to understand the historical course of the Viking expansion across the North Atlantic. Without adhering to geographical determinism, one cannot deny the influence of physical geography on the possibilities of a human population to travel to and survive on an island. A remote island is hard to reach, not only for humans, but for most animals and plants as well. This leads to a reduced biodiversity and consequently a lower carrying capacity on more remote islands. The same is true for smaller islands, where only small populations of a limited number of species can be sustained. Thus, generally speaking, larger and less remote islands can sustain larger human populations (Keegan & Diamond 1987: 58-60, 62-65). It can reasonably be assumed that accessibility and carrying capacity partly define the attractiveness of an island. In other words, all else being equal, potential colonists will prefer a large island to a small one, and a close one to a remote one. By quantifying these characteristics, a relative ranking of the attractiveness of islands can be constructed. -8- 8/01/2007 h:01 a1/p1 Area (table 1) In the case of the barren islands of the North Atlantic, it is obvious that useful and not total surface area is the relevant measure for carrying capacity. This area is defined in table 1 as the area suitable for livestock farming, the only terrestrial subsistence activity that was exported more or less unaltered from Norway to every island. Maximal (100%) suitability has been used in this table in lack of more precise data for Orkney, Shetland and the Faeroes. This should be interpreted as the absence of obvious, large stretches of unsuitable land. According to Eysteinsson and Blöndal (2003: 411) the original forest coverage of Iceland was 27%, and these areas were potential pasture. UNSD reports 22% land presently under pastures in Iceland, which, given the significant landscape degradation since landnám, confirms the above assumption. Vínland is excluded from the table because the aim of the expeditions there was to collect lumber and high-value low-bulk goods such as hides and vines, not to find suitable land for settlement. Moreover, it is impossible to determine what area should be used in the analysis: the island of Newfoundland was only the location of the base camp, not the exploited region, whereas the extent of Vínland is unknown. Each island/island group has been ranked according to the criterion of suitability for livestock farming (column 6 of table 1), revealing the great attractiveness of Iceland and (perhaps surprisingly) Greenland to Norse pastoralists. Distance (tables 2 and 3) Table 2 contains the direct, Euclidean distances between the different islands and island groups in the North Atlantic, as measured with the Google Earth measuring tool. Because it is known historically that Greenland was colonized from Iceland, and because Iceland is always ‘in the way’ on direct routes from the other islands to Greenland (implying that it was an automatic stepping stone), only the distance Iceland-Greenland is relevant to this analysis. Like area, the analysis of distance is not straightforward. There is no doubt about the capacity of the Norse to travel directly from Norway to the settlements in Greenland (Magnusson & Pálsson 1965: 16; Christensen 2000: 95). The sea voyages of the earliest settlers were of a different nature, however. Livestock, especially cattle, can be extremely hard to handle at sea, and regular stops must have been included in the travel route to feed and drench the animals (Broodbank & Strasser 1991: 240). On later voyages, when landfall was not necessary, the islands were used as navigational points of reference. Therefore, not only the total distance from point of origin to destination has to be taken into account, but also the so-called ‘longest single voyage’ (LSV), the longest distance covered between stops (Patton 1996: 40). Obviously, this assumed tendency to minimize the time spent on open sea has implications for the choice of the staging point of the voyages. Earlier on, Norway has been indicated as the country of origin of many of the settlers. It seems reasonable to assume that the first settlers, regardless of their exact place of origin within Norway, kept close to land as long as possible, sailing the so-called Northway along the sheltered Norwegian coast. The ideal starting point would thus be the regions Hordaland and Sogn og Fjordane in the southwest of Norway, where the distance to the closest island group, Shetland, is only about 320 km. Each of the islands is connected to the others by a number of possible itineraries, using different stepping stones. It can be assumed that the first colonists selected those routes that maximized the difference between direct distance and LSV, while at the same time minimizing the travel time (i.e. minimizing the difference between direct distance and the total sum of the distances between stepping stones). The most plausible of these travel routes according to this reasoning, their distances and the stepping stones used are shown in table 3. Note that Fair Isle 2, approximately midway between Orkney and Shetland, is not involved in this analysis, as 2 Indications for Viking presence on Fair Isle are scarce. The only sources are the Norse toponymy of the island, and a handful of mentions in the Orkneyinga Saga. Nonetheless, it is clear that the island was, as it is now, an important beacon on the route between Orkney and Shetland (the name of Fair Isle may derive from ‘Fire Island’) (Hunter 1996: 103-115; Graham-Campbell & Batey 2001: 67). -9- 8/01/2007 h:01 a1/p1 breaking up the already small distance between Orkney and Shetland with another stepping stone would not have any significant effect on the relative ranking of the North Atlantic islands. As can been seen in table 3, Shetland and above all the Faroe Islands fulfilled a crucial role as stepping stones for the first Norse settlers in the North Atlantic. The use of stepping stones splits the North Atlantic into short ‘steps’ that never exceed 500 km. Experiments with full- scale reconstructions of Viking-period ships (e.g. Damgård-Sørensen 2002: 225-227) have shown that this is exactly the distance that could be covered in a few days. Attractiveness (table 4) The standard way of calculating the index of attractiveness (IA) is by dividing area by distance (fig. 2a) (Cherry 1985: 17). Either total distance or LSV can be used as a denominator in this formula. However, it can be assumed that both LSV and total distance affect attractiveness simultaneously. Therefore, a formula that incorporates both (fig. 2b) is used here to calculate the IA of the North Atlantic islands (table 4). Subsequently, the islands were ranked according to their attractiveness. From a Norwegian point of view (i.e. excluding Greenland), Iceland is without doubt the most attractive area for colonization, followed by Shetland. Orkney and the Faroe Islands follow and have a very similar score. A next step in this analysis will be to compare the results of this analysis with the chronological course of the Norse colonization. Any deviations from this pattern will need to be explained by in terms of push factors in the homeland, or pull factors in the islands other than suitable land for livestock tending. The stages of island colonization From a biogeographical point of view, island colonization is a single, unambiguous event: the arrival of one or more individuals of a certain plant or animal species on an island. It is obvious that island colonization by humans is a much more complex phenomenon comprising many activities and events. In the light of this understanding, colonization must be seen as a process divisible into separate phases that always occur in the same order (Graves & Addison 1995: 386-387; Guerrero 2001: 139-140). The first phase consists of discovery and exploration. During these activities, knowledge of the position, geography and economic assets of the island is obtained, and the routes to the islands are explored (Graves & Addison 1995: 386-387; Guerrero 2001: 139, 141). If we assume that Dicuil and other writers were correct in stating that Faeroe and Iceland were visited by Irish Christians in the eighth century, it is possible that the Vikings learned of the position of these islands during their early visits to the British Isles. However, there is no evidence to confirm this. Later written sources state that Iceland, Greenland and Vínland were discovered coincidentally, often by sailors driven off course. In other words, the Norse did not actively search for new lands to settle. This is in contrast with the autocatalytic models proposed for the colonization of, for example, the Pacific, where the discovery of new islands triggered even more exploration (Keegan & Diamond 1987: 67; Irwin 1990: 91-92). Often, the exploration of an island takes the shape of incorporation and exploitation of the island in economic networks (e.g. Cherry 1985: 20; Patton 1996: 83; Broodbank 2000: 127-129; all pointing out episodes of economic exploitation prior to island colonization in the Aegean). Several examples from the North Atlantic have already been mentioned. Orkney and possibly Shetland may have taken part in trade with Scandinavia well before the first Norse settled on the islands. Vínland is a special case, in that it provides archaeological and literary evidence of the exploration stage, which is otherwise rarely documented. These sources of information elucidate several other crucial points about this stage. Firstly, the expeditions had their own goals, in this case economic exploitation. In other words, they were not necessarily organised with assessment of the potential for permanent settlement as their main objective (Anthony - 10 - 8/01/2007 h:01 a1/p1 1990: 903). Secondly, these differing goals were reflected in the composition of the expedition (Wallace 2003a: 214). While a colonist group would constitute a complete social group, with men, women and children, the members of an expedition are almost exclusively young men, some specifically selected because of their experience or specialisation in tracking and hunting, ship-repairing, etc. This difference in the composition of groups aimed at exploration versus colonization is another fundamental feature of any migration process (Anthony 1990: 905). The result of the exploration stage is a (subjective) assessment of the potential of the island to sustain a population. This knowledge of the carrying capacity and other features of an island is a crucial element in the decision-making process prior to and the preparation for a colonization attempt. In a second stage, colonization proper, permanent settlements inhabited by a complete social group are founded. This group of settlers now faces a demographic problem – the survival of a small group; the so-called beachhead bottleneck – and an economic-ecological problem – the survival in a new natural environment (Keegan & Diamond 1987: 74). When preparing to settle on an uninhabited island, anticipating these problems greatly improves the chances of success. Discovery of an island may be coincidental, colonization rarely is. The survival of a small group of settlers primarily depends on preparations made before the expedition, especially regarding food stocks and composition of the group (Broodbank & Strasser 1991: 239-240). However, sheer luck is a factor that should not be underestimated: computer simulations (e.g. McArthur 1976) show that stochastic oscillations determine the chance of survival of small populations to a large degree. The second problem can be tackled by adapting economic strategies, but also by changing the natural environment to better suit the requirements of the settlers. Possibilities includes the importation of new plant and animal species, deforestation, etc. (Keegan & Diamond 1987: 74, 77-78; Graves & Addison 1995: 386-387). This is by no means an exclusively economic process, but also implies imposing mental categories or ‘transported landscapes’ (Kirch 1982) on an alien environment. Owing to the apparent similarity of the new landscape with that of their homeland, Norse settlers were able to transport their traditional pastoralist economy to the islands without having to change it significantly. Only later, partly due to climate changes, it became clear that in the long term this economy had detrimental effects on the landscape, particularly in Iceland (McGovern et al. 1988; Buckland et al. 1991: 259-269; Amorosi et al. 1997; Simpson et al. 2001; Eysteinsson and Blöndal 2003: 413-414). This failure to adapt eventually led to the disappearing of the Norse population in Greenland (McGovern 1980, 2000). The traditional accounts of Norse expansion across the North Atlantic provide some illustrations of these problems. For instance, according to Landnámabók, the first attempt to colonise Iceland by Floki Vilgerdarson in 865 failed because he spent his summer hunting and fishing instead of collecting hay. Consequently, his cattle starved during the first winter and Floki was forced to return to Norway (Adalsteinsson 1991: 291). Floki undoubtedly considered his failed colonization attempt a debacle – after all, he was the one who allegedly named the new land Iceland. For the later settlers, who started arriving in large numbers a decade later, Floki’s stories (regardless of whether they were true or not) prevented them making the same mistakes. Thus, this failed colonization attempt had a similar function as the earlier mentioned economic exploitation of Orkney or Vínland: gathering vital information to improve the survival chances of later settlers. The final stage of island colonization mentioned in literature is settlement. In this stage, the population reaches a number high enough to ensure long-term survival (Keegan & Diamond 1987: 74-78; Graves & Addison 1995: 386-387). The minimum number of individuals required to reach biological autonomy is often said to be 500 (Jones 1976: 364). It is obvious, however, - 11 - 8/01/2007 h:01 a1/p1 that this last stage is not only very hard to recognize in the archaeological record, but not even relevant in the light of the high mobility of the Vikings. Chronology of the North Atlantic colonization Table 5 recapitulates the approximate dates for discovery and colonization of the different islands, as established in the first section of this paper. As mentioned above, most of these dates are problematic and subject to debate. The dates marked with an asterisk (*) derive from written sources but are acceptable in light of the archaeological evidence. The unmarked dates are based solely on the interpretation of the archaeological record and should be considered rough and inherently uncertain estimates. In the cases of Iceland and Newfoundland/Vínland, the colonization date given refers to a failed colonization attempt known from literature. This is because the fact that someone found the island attractive enough to settle is of interest here, not the success of the attempt. To eliminate the effect of distance, a new relative ranking has been worked out based on the duration of the exploration stage, i.e. the period of time between discovery and (attempted) colonization (columns 3 and 4). It is reasonable to assume that the shorter the exploration stage, the more attractive the island was for potential colonists. As such, this chronological ranking provides an alternative to the biogeographical index of attractiveness, and the confrontation between the two (table 6) potentially sheds some light on the motivations of the settlers. There are some similarities between the two rankings, such as the dominant position of Iceland and the intermediate positions of the Faroes and Shetland. The main differences are the position of Greenland, which has a high biogeographical attractiveness but did not attract any settlers until decades after it was first discovered, and Orkney, that scores low in the biogeographical ranking, but was settled at the same time as the (seemingly) more attractive Shetland. These discrepancies need to be explained by motivational factors other than available land for livestock tending. The three-stage model In an attempt to explain these differences between chronological and biogeographical rankings, and taking into account the chronological course and historical background of the expansion, a three-stage model for the Viking colonization of the North Atlantic islands is proposed. 1. During stage 1 (seventh or eight to mid-ninth centuries AD) Orkney, Shetland and the Faroes were discovered, explored and, in the case of the former two, exploited (through trade and possibly looting). The first Viking settlers arrived, albeit in relatively limited numbers. Orkney was attractive as a gateway to the riches of the British Isles, and had more agricultural potential than any of the other North Atlantic Islands. Shetland had its own Pictish riches that could be traded and looted and possibly functioned as a stepping stone to the British Isles. The attracting factors of the Faroes are less clear, and the motivation for settlers (if any arrived at all in this early phase) were most likely of a more individual nature, such as social exclusion. 2. In the final decades of the ninth century AD, the colonization process accelerated greatly. The most likely explanation for this is the turbulent period Norway was going through at this time, generating large numbers of fugitives. These newcomers triggered socio-political reorganisation in Orkney. In Shetland and Orkney, this new stream of immigrants may have been the cause of the ultimate disappearing of the Pictish language and culture. In this context of increased population movement, maritime activity around the islands increased, so not surprisingly Iceland was discovered very quickly. Shortly after this discovery, numerous settlers travelled to the new land. Its location makes the Faroes an ideal stepping stone to Iceland, and as such these islands may have become significantly - 12 - 8/01/2007 h:01 a1/p1  more attractive in this period. Iceland, because of its size, acted as a ‘settler trap’. The discovery of Greenland therefore did not trigger immediate exploration (let alone colonization) of potential new living areas, simply because no one was interested. 3. The final stage consists of the exploration and colonization of Greenland and the exploitation and attempted colonization of Vínland. Shortly before 1000 AD, a group of Icelandic malcontents lead by the charismatic Eirik the Red settled in Greenland, which at that time was known for half a century. By coincidence, a new and rich land was discovered even further to the west. At first the resources of Vínland were exploited because of wealth and prestige they brought to the members of the expeditions. Perhaps captured by a spirit or mentality of migration, or a colonization ideology, the Greenland Norse tried to settle in Vínland, but this attempt was doomed to failure. This model makes clear that the Viking colonization of the North Atlantic islands was not a simple process of westward expansion, and that it cannot be sufficiently explained by a pointing to a single cause or motive. At different times, the colonization process was characterised by a particular speed, scale and destination. Particularly notable are the different combinations of push factors in the homeland and pull factors in the islands that fuelled the expansion through time. Conclusion The above paragraphs demonstrated that an island archaeological approach is able to offer fresh insights, even concerning a migration episode as well attested archaeologically and historically as the Viking expansion in the North Atlantic. The understanding of island colonization helps to make sense of the available information, and to identify gaps in our current knowledge. By combining biogeographical analysis and theories on colonization with the preserved sources of evidence, separate phases of expansion can be identified, each characterised by a particular time, origin, destination, cause and scale. This leads to a broad perspective of the Viking expansion in the North Atlantic as a whole. As such, this is a plea for an island archaeology of the North Atlantic, which certainly possesses the same potential for this approach as the more trodden paths of the Mediterranean and Pacific. As a conclusion, a few areas are suggested in which this research can be expanded, in order to gain a deeper understanding of the process of the Viking migrations across the North Atlantic. This, in turn, may have implications for our knowledge of the nature of the Viking expansion in general. The first, and most obvious, possibility of continuing this research is to expand the geographical area of study, at least to include the northern coasts of Scotland and the Hebrides, which were well integrated in this North Atlantic ‘island world’. This leads to a second, closely related, potential extension to the present study. In this paper, only the southwest of Norway is used as a staging area for the voyages of the Viking explorers and settlers. However, as genetic research increasingly points out, parts of the British Isles were not only settled by the Norse, but also contributed significantly to the settler groups of other islands, particularly the Faroes and Iceland. Therefore, it would be most useful (but also complicating) to view the colonization of the latter islands from a second area of origin, e.g. at the northernmost end of the Isle of Lewis or the northwestern point of mainland Scotland. A final suggestion for further research concerns the biogeographical methods employed. Only a relatively simple calculation involving area and distance has been used in this paper. The use of several other geographical features is proposed in island archaeological literature, sometimes related to a particular stage in the colonization process. For instance, the so-called T/D ratio or target-distance ratio expresses the angle that an island occupies on the horizon of the staging point in function of the distance from staging point to the island, and is used in estimating the chance of discovery (Held 1989: 13; Patton 1996: 40-41). Applied to the North Atlantic, this - 13 - 8/01/2007 h:01 a1/p1 method could for instance contribute to the question of the role of Irish geographical knowledge in the Viking discoveries of the Faroes and Iceland. Acknowledgments I would like to thank my promoter, Marc Lodewijckx (Catholic University of Leuven), for his guidance and support during the writing of my licentiate thesis, part of which forms the basis of this paper. I am also very grateful to James Barrett (University of York) for his suggestions on an earlier draft of this paper. Any errors remain, of course, entirely my own responsibility. References S. Adalsteinsson (1991) Importance of sheep in early Icelandic agriculture, Acta Archaeologica 61, 285-291. B. Agozino (ed.) (2000) Theoretical and Methodological Issues in Migration Research. Interdisciplinary, intergenerational and international perspectives, Aldershot. K. Ahronson (2000) Further evidence for a Columban Iceland: Preliminary results of recent work, Norwegian Archaeological Review 33: 117-124. K. Ahronson (2002) Testing the evidence for northern North Atlantic Papar: A cave site in Southern Iceland, in: B.E. Crawford (ed.) The Papar in the North Atlantic: Environment and History. The ‘Papar’ Project vol. 1 (St John’s House Papers, No. 10), St Andrews: 107-120. K. Ahronson (2003) The crosses of Columban Iceland: A survey of preliminary research, in: S. Lewis-Simpson (ed.) Vínland Revisited: The Norse World at the Turn of the First Millennium. Selected Papers from the Viking Millennium International Symposium, 15-24 September 2000, Newfoundland and Labrador, St John's: 75-82. S.E. Albrethsen (2003) The early Norse farm buildings in Western Greenland: Archaeological evidence, in: S. Lewis-Simpson (ed.) Vínland Revisited: the Norse World at the Turn of the First Millennium. Selected Papers from the Viking Millennium International Symposium, 15-24 September 2000, Newfoundland and Labrador, St John's: 97-110. S. Amin (1995) Migrations in contemporary Africa: A retrospective view, in: J. Baker, and T.A. Ina (eds.) The migration experience in Africa, Stockholm: 29-40. T. Amorosi, P. Buckland, A Dugmore, J.H. Ingimundarson, and T.H. McGovern (1997) Raiding the landscape: Human impact in the Scandinavian North Atlantic, Human Ecology 25: 491-518. D. Anthony (1990) Migration in archaeology: The baby and the bathwater, American Anthropologist 92(4): 895-914. S.V. Arge (1993) On the landnam of the Faroe Islands, in: C.E. Batey, J. Jesch, and C.D. Morris (eds.) The Viking Age in Caithness, Orkney and the North Atlantic, Edinburgh: 465-472. S.V. Arge, and N. Hartmann (1992) The Burial Site of við Kirkjugarð in the Village of Sandur, Sandoy, Fróðskaparrit 38-39: 5-21. S.V. Arge, G. Sveinbjarnardóttir, K.J. Edwards, P.C. Buckland (2005) Viking and Medieval settlement in the Faroes: People, place and environment, Human Ecology 33: 597-620. J. Arneborg (2000) Greenland and Europe, in: W.W. Fitzhugh, and E.I. Ward (eds.) Vikings. The North Atlantic Saga, Washington-London: 304-317. J. Arneborg (2003) Norse Greenland: Reflections on settlement and depopulation, in: J.H. Barrett (ed.) Contact, Continuity and Collapse. The Norse Colonization of the North Atlantic, Turnhout: 163-181. J. Arneborg, J. Heinemeier, N. Lynnerup, H.L. Nielsen, en Á.E. Sveinbjörnsdóttir (1999) Changes of diet of the Greenland Vikings determined from stable carbon isotope analysis and 14C dating of their bones, Radiocarbon 41: 157-168. - 14 - 8/01/2007 h:01 a1/p1 J.H. Barrett (2003) Culture contact in Viking Age Scotland, in: J.H. Barrett (ed.) Contact, Continuity, and Collapse. The Norse Colonization of the North Atlantic (Studies in Early Middle Ages 5), Turnhout: 73-111. B. Bass (1998) Early Neolithic offshore accounts: Remote islands, maritime exploitations, and the trans-Adriatic cultural network, Journal of Mediterranean Archaeology 11: 165-190. Z. Bauman (1995) Life in Fragments: Essays in Post-Modern Morality, Oxford. J. Berglund (2000) The farm beneath the sand, in: W.W. Fitzhugh, and E.I. Ward (eds.) Vikings. The North Atlantic Saga, Washington-Londen: 295-303. S. Bowdler (1995) Offshore islands and maritime explorations in Australian prehistory, Antiquity 69: 945-958. R. Boyer (1992) Les Vikings. Histoire et civilisation, Paris. C.B. Brettell, and J.F. Hollifield (2000) Migration Theory: Talking Across Disciplines, New York-London. C. Broodbank (1999) Colonization and configuration in the insular Neolithic of the Aegean, in: P. Halstead (ed.) Neolithic Society in Greece (Sheffield Studies in Aegean Archaeology 2), Sheffield: 15-41. C. Broodbank (2000) An Island Archaeology of the Early Cyclades, Cambridge. C. Broodbank, and T.F. Strasser (1991) Migrant farmers and the Neolithic colonization of Crete, Antiquity 65: 233-245. P.C. Buckland, A. Dugmore, D.W. Savory, and G. Sveinbjarnardóttir (1991) Holt in Eyjafjallasveit, Iceland: A paleoecological study of the impact of Landnám, Acta Archaeologica 61: 252-271. S. Burmeister (2000) Archaeology and migration: Approaches to an archaeological proof of migration, Current Anthropology 41: 539-568. J. Chapman, en H. Hamerow (1997) Migrations and Invasions in Archaeological Explanation (BAR International Series 664), Oxford. J.F. Cherry (1981) Pattern and process in the earliest colonization of the Mediterranean islands, Proceedings of the Prehistoric Society 47: 41-68. J.F. Cherry (1985) Islands out of the stream: Isolation and interaction in early East Mediterranean insular prehistory, in: A.B. Knapp, and T. Stetch (eds.) Prehistoric Production and Exchange: the Aegean and Eastern Mediterranean (Monograph XXV), Los Angeles: 12-29. B. Crawford (1987) Scotland in the Early Middle Ages, 2. Scandinavian Scotland, Leicester. A.E. Christensen (2000) Ships and navigation, in: W.W. Fitzhugh, and E.I. Ward (eds.) Vikings. The North Atlantic Saga, Washington-London: 86-97. S. Dahl (1965) A survey of archaeological investigations in the Faeroes, in: A. Small (ed.) Fourth Viking Congress, Aberdeen: 135-141. T. Damgård-Sørensen (2002) Les bateaux de Skuldelev (Roskilde) et leurs répliques modernes, in: É Ridel (ed.) L’héritage maritime des Vikings and Europe de l’Ouest, Caen: 199-228. H.J. Debes (1993) Problems concerning the earliest settlement in the the Faroes Islands, in: C.E. Batey, J. Jesch, and C.D. Morris (eds.) The Viking Age in Caithnes, Orkney and the North Atlantic, Edinburgh: 454-464. P. Deckers (2005) Eilandarcheologie. Een overzicht van de theorie and een casestudy over de Vikingexpansie in de Noordelijke Atlantische Oceaan (unpublished licentiate thesis, K.U. Leuven). G. Dorigo and W. Tobler (1983) Push-pull migration laws, Annals of the Association of American Geographers 73: 1-17. A.J. Dugmore, M.J. Church, P.C. Buckland, K.J. Edwards, I. Lawson, T.H. McGovern, E. Panagiotakopulu, I.A. Simpson, P. Skidmore and G. Sveinbjarnardóttir (2005) The Norse - 15 - 8/01/2007 h:01 a1/p1  landnám on the North Atlantic islands: An environmental impact assessment, Polar Record 41 (216): 21-37. K.J. Edwards (2005) "On the windy edge of nothing”: A historical human ecology of the Faroe Islands, Human Ecology 33: 585-596. Þ. Eysteinsson, en S. Blöndal (2003) The forests of Iceland at the time of settlement: Their utilisation and eventual fate, in: S. Lewis-Simpson (ed.) Vínland Revisited: the Norse World at the Turn of the First Millennium. Selected Papers from the Viking Millennium International Symposium, 15-24 September 2000, Newfoundland and Labrador, St John's: 411-415. A. Fenton (1997) The Northern Isles: Orkney and Shetland, Phantassie. P. Foote (1970) On legal terms in Færeyinga Saga, Fróðskaparit 18: 159-175. A. Friðriksson and O. Vésteinsson (2003) Creating a past: A historiography of the settlement of Iceland, in: J.H. Barrett (ed.) Contact, Continuity and Collapse. The Norse Colonization of the North Atlantic, Turnhout: 139-161. S. Goodacre, A. Helgason, J. Nicholson, L. Southam, L. Ferguson, E. Hickey, E. Vega, K. Stefánsson, R. Ward, and B. Sykes (2005) Genetic evidence for a family-based Scandinavian settlement of Shetland and Orkney during the Viking periods, Heredity 95: 129-135. J. Graham-Campbell (2001) The Viking World, Londen. J. Graham-Campbell, and C.E. Batey (2001) Vikings in Scotland. An Archaeological Survey, Edinburgh. M.W. Graves, and D.J. Addison (1995) The Polynesian settlement of the Hawaiian Archipelago: Integrating models and methods in archaeological interpretation, World Archaeology 26: 380-399. K. Grönvold, K. Óskarsson, S.J. Johnsen, H.B. Clausen, C.U. Hammer, G. Bond, and E. Bard (1995) Ash layers from Iceland in the Greenland GRIP ice core correlated with oceanic and land sediments, Earth and Planetary Science Letters 135: 149–55. V.M. Guerrero Ayuso (2001) The Balearic Islands: Prehistoric colonization of the furthest Mediterranean islands from the mainland, Journal of Mediterranean Archaeology 14: 136- 158. H.C. Gulløv (2000) Natives and Norse in Greenland, in: W.W. Fitzhugh, and E.I. Ward (eds.) Vikings. The North Atlantic Saga, Washington-Londen: 318-326. E.O. Halldórsson (1961) Um landnám Gríms Kambans í Føroyum, Fróðskaparit 10: 47-52. L. Hedeager (2000) From warrior to trade economy, in: W.W. Fitzhugh, and E.I. Ward (eds.) Vikings. The North Atlantic Saga, Washington-Londen: 85-86. S.O. Held (1989) Colonization cycles on Cyprus, 1: The biogeographic and palaeontological foundations of early prehistoric settlement, Report of the Department of Antiquities, Cyprus 1989: 7-28. A. Helgason, E. Hickey, S. Goodacre, J. Nicholson, B. Sykes, E.W. Hill, D.G. Bradley, V. Bosnes, J.R. Gulcher, R. Ward, and K. Stefánsson (2000a) Estimating the Proportions of Norse and Gaelic Ancestry in the Male Settlers of Iceland, American Journal of Human Genetics 67: 697-717. A. Helgason, S. Sigurðadóttir, J.R. Gulcher, R. Ward, and K. Stefánsson (2000b) mtDNA and the origin of the Icelanders: Deciphering signals of recent population history, American Journal of Human Genetics 66: 999-1016.. A. Helgason, E. Hickey, S. Goodacre, V. Bosnes, K. Stefánsson, R. Ward and B. Sykes (2001) mtDNA and the islands of the North Atlantic: Estimating the proportions of Norse and Gaelic ancestry, American Journal of Human Genetics 68: 723-737. - 16 - 8/01/2007 h:01 a1/p1 K. Helle (1998) The history of the Early Viking Age in Norway, in: H. Clarke, M. Ní Mhaonaigh, and R. Ó Floinn (eds.) Ireland and Scandinavia in the Early Viking Age, Dublin: 239-258. J.R. Hunter (1996) Fair Isle. The Archaeology of an Island Community, Edinburgh. J.R. Hunter, J.M. Bond, and A.N. Smith (1993) Some aspects of early Viking settlement in Orkney, in: C.E. Batey, J. Jesch, and C.D. Morris (eds.) The Viking Age in Caithness, Orkney and the North Atlantic, Edinburgh: 272-284. G.J. Irwin (1996) The Prehistoric Exploration and Colonization of the Pacific, Cambridge. A. Jackson (1991) The Faroes. The Faraway Islands, London. G. Jones (1999) ‘The same people in the same places’? Socio-spatial identities and migration in youth, Sociology 33: 1-22. R. Jones (1976) Man as an element of a continental fauna: The case of the sundering of the Bassian bridge, in: J. Golson, R. Jones, and J. Allen (eds.) (1976) Sunda and Sahul, Londen: 317-386. T.H. Jorgensen, H.N. Buttenschön, A.G. Wang, T.D. Als, A.D. Børglum, and H. Ewald (2004) The origin of the isolated population of the Faroe Islands investigated using Y chromosomal markers, Human Genetics 115: 19-28. W.F. Keegan, and J.M. Diamond (1987) Colonization of islands by humans: A biogeographical perspective, in: M.B. Schiffer (ed.) Advances in Archaeological Method and Theory 10, San Diego-New York: 49-92. C. Keller (1991) Vikings in the West Atlantic: A model of Norse Greenlandic Medieval Society, Acta Archaeologica 61, 126-141. P.V. Kirch (1982) Transported landscapes, Natural History 91(12) 32-34. L.O. Lagerqvist (1992) Les monnaies scandinaves, in: E. Roesdahl, J.-P. Mohen, and F.-X. Dillmann (eds.) Les Vikings… Les Scandinaves et l'Europe, 800-1200, Paris: 220-221. E.S. Lee (1966) A theory of migration, Demography 3: 37-57. K. Lunden, B. Myhre, J. Sandnes and H. Steuer (2001) Comments on the social context of settlement in Norway in the First Millennium AD, Norwegian Archaeological Review 34: 13-23. N. Lynnerup (1996) Paleodemography of the Greenland Norse, Arctic Anthropology 33.2: 122- 136. N. Lynnerup (2000) Life and death in Norse Greenland, in: W.W. Fitzhugh, and E.I. Ward (eds.) Vikings. The North Atlantic Saga, Washington-Londen: 285-294. L. MacGregor (1984) Sources for a study of Norse settlement in Shetland and the Faroes, in: B.E. Crawford (ed.) Essays in Shetland History, Lerwick: 1-17. M. Magnusson, en P. Pálsson (1965) The Vinland Sagas. The Norse Discovery of America. Groenlendinga Saga and Eirik’s Saga, Harmondsworth. D.L.D. Mahler (1991) Argisbrekka: New evidence of shielings in the the Faroes Islands, Acta Archaeologica 61, 60-72. D.L.D. Mahler (1993) Shielings and their role in the Viking-Age Economy, in: C.E. Batey, J. Jesch, and C.D. Morris (eds.) The Viking Age in Caithness, Orkney and the North Atlantic, Edinburgh: 487-505. G. Malmberg (1997) Time and space in international migration, in: T. Hammar, G. Brochmann, K. Tamas, en T. Faist (eds.) International Migration, Immobility and Development. Multidisciplinary Perspectives, Oxford-New York: 21-48. N. McArthur (1976) Computer simulations of small populations, Australian Archaeology 4: 53- 57. T.H. McGovern (1980) Cows, harp seals, and churchbells: Adaptation and extinction in Norse Greenland, Human Ecology 8(3), 245-275. - 17 - 8/01/2007 h:01 a1/p1 T.H. McGovern (2000) The demise of Norse Greenland, in: W.W. Fitzhugh, and E.I. Ward (eds.) Vikings. The North Atlantic Saga, Washington-Londen: 327-339. T.H. McGovern, G. Bigelow, T. Amorosi, and D. Russell (1988) Northern islands, human error, and environmental degradations: a view of social and ecological change in the medieval North Atlantic, Human Ecology 16: 225-270. S. McGrail (2001) Boats of the World: From the Stone Age to Medieval Times, Oxford. C.D. Morris (1985) Viking Orkney: A Survey, in: C. Renfrew (ed.) The Prehistory of Orkney, Edinburgh: 210-242. B. Myhre (1993) The beginning of the Viking Age – Some current archaeological problems, in: A. Faulkes, and R. Perkins (eds.) Viking Revaluations. Viking Society Centenary Symposium, London: 182-204. B. Myrhe (1998) The archaeology of the Early Viking Age in Norway, in: H. Clarke, M. Ní Mhaonaigh, and R. Ó Floinn (eds.) Ireland and Scandinavia in the Early Viking Age, Dublin: 3-36. B. Myhre (2000) The Early Viking Age in Norway, in: S. Stummann Hansen, and K. Randsborg (eds.) Vikings in the West (Acta Archaeologica 71), Kopenhagen: 35-47. U. Näsman (2000) Raids, migrations, and kingdoms - The Danish case, in: S. Stummann Hansen, and K. Randsborg (eds.) Vikings in the West (Acta Archaeologica 71), Kopenhagen: 1-7. D. Omand (1993) The landscape of Caithness and Orkney, in: C.E. Batey, J. Jesch, and C.D. Morris (eds.) The Viking Age in Caithness, Orkney and the North Atlantic, Edinburgh: 102- 110. O. Owen (2002) Les Vikings en Écosse: quel type de maison les colons vikings construisaient- ils?, in: É Ridel (ed.) L'héritage maritime des Vikings and Europe de l'Ouest, Caen: 137- 170. O. Owen (2004) The Scar boat burial - and the missing decades of the Early Viking Age in Orkney and Shetland, in: J. Adams, and K. Holman (eds.) Scandinavia and Europe 800- 1350: Contact, Conflict and Co-existence (Medieval Texts and Cultures of Northern Europe, 4), Turnhout: 3-33. H. Pálsson, and P. Edwards (1981) Orkneyinga Saga. The History of the Earls of Orkney, Londen-New York. M. Patton (1996) Islands in Time. Island Sociogeography and Mediterranean Prehistory, London. E. Roesdahl (1992) Les royaumes scandinaves, in: E. Roesdahl, J.-P. Mohen, and F.-X. Dillmann (eds.) Les Vikings… Les Scandinaves et l'Europe, 800-1200, Paris: 32-41. E. Roesdahl (2003) The Scandinavians at home, in: D.M. Wilson (ed.) The Northern World, Londen: 127-158. J. Saillard, P. Forster, N. Lynnerup, H.-J. Bandelt, and S. Nørby (2000) mtDNA variation among Greenland Eskimos: The edge of the Beringian expansion, American Journal of Human Genetics 67: 718-726. P. Sawyer (2003) Scotland, Ireland and Iceland: Norwegian settlers in the ninth century, in: S. Lewis-Simpson (ed.) Vínland Revisited: the Norse World at the Turn of the First Millennium. Selected Papers from the Viking Millennium International Symposium, 15-24 September 2000, Newfoundland and Labrador, St John's: 29-35. P. Schledermann and K.M. McCullough (2003) Inuit-Norse Contact in the Smith Sound Region, in: J.H. Barrett (ed.) Contact, Continuity and Collapse. The Norse Colonization of the North Atlantic, Turnhout: 183-205. S.H. Schneider, en R. Londer (1984) The Co-Evolution of Climate and Life, San Francisco. K.A. Seaver (2000) Unanswered questions, in: W.W. Fitzhugh, and E.I. Ward (eds.) Vikings. The North Atlantic Saga, Washington-Londen: 270-279. - 18 - 8/01/2007 h:01 a1/p1 H. Shetelig (1954) The Viking graves, in H. Shetelig (ed.) Viking Antiquities in Great Britain and Ireland IV, Oslo: 65-111. I.A. Simpson, A.J. Dugmore, A. Thomson, and O. Vésteinsson (2001) Crossing the tresholds: Human ecology and historical patterns of landscape degradation, Catena 42: 175-192. D. Skre (2001a) The social context of settlement in Norway in the first millennium AD, Norwegian Archaeological Review 34: 1-12 D. Skre (2001b) Reply to Comments on the social context of settlement in Norway in the first millennium AD, Norwegian Archaeological Review 34: 24-34. A. Small (1969) The distribution of settlement in Shetland and the Faroes in Viking times, Saga Book of the Viking Society 17 (1967-1968): 145-155. K.P. Smith (1995) Landnám: The settlement of Iceland in archaeological and historical perspective, World Archaeology 26: 319-347. J. Steffensen (1975) Menning og meinsemdir: Rigerðarsafn um mótunarsögu íslenskar pjóðar og baráttu hennar við hungur og sóttir, Reykjavik. S. Stummann Hansen (2000) Viking settlement in Shetland. Chronological and regional contexts, in: S.Stummann Hansen, and K. Randsborg (eds.) Vikings in the West (Acta Archaeologica 71), Kopenhagen: 87-103. S. Stummann Hansen (2003) The early settlement of the Faroe Islands: The creation of cultural identity, in: J.H. Barrett (ed.) Contact, Continuity and Collapse. The Norse Colonization of the North Atlantic, Turnhout: 33-71. G. Sveinbjarnardóttir (2002) The question of papar in Iceland. In B.E. Crawford (ed.) The Papar in the North Atlantic: Environment and History. The ‘Papar’ Project vol. 1 (St John’s House Papers, No. 10), St Andrews: 97-106. H. Thorláksson (2000) The Icelandic Commonwealth Period. Building a new society, in: W.W. Fitzhugh, and E.I. Ward (eds.) Vikings. The North Atlantic Saga, Washington-Londen: 175- 185. V. Turner (1998) Ancient Shetland, London. V. Turner, S. Dockrill, en J. Bond (2005) Viking settlement in an Iron Age village: Old Scatness, Shetland, in: A. Mortensen, en S.V. Arge (eds.) Viking and Norse in the North Atlantic. Select Papers from the Proceedings of the Fourteenth Viking Congress, Tórshavn, 19-30 July 2001 (Annales Societatis Scientiarum Færoensis Supplementum XLIV), Tórshavn: 245-249. O. Vésteinsson (2000) The archaeology of landnám. Early settlement in Iceland, in: W.W. Fitzhugh, and E.I. Ward (eds.) Vikings. The North Atlantic Saga, Washington-Londen: 164- 174. F.T. Wainwright (1962) The Scandinavian settlement, in: F.T. Wainwright (ed.) The Northern Isles, Edinburgh: 117-162. B.L. Wallace (1991) L'Anse aux Meadows. Gateway to Vinland, Acta Archaeologica 61, 166- 197. B.L. Wallace (2000) The Viking settlement at L'Anse aux Meadows, in: W.W. Fitzhugh, and E.I. Ward (eds.) Vikings. The North Atlantic Saga, Washington-Londen: 208-216. B.L. Wallace (2003a) L’Anse aux Meadows and Vinland: An abandoned experiment, in: J.H. Barrett (ed.) Contact, Continuity and Collapse. The Norse Colonization of the North Atlantic, Turnhout: 207-238. B.L. Wallace (2003b) The later excavations at L'Anse aux Meadows, in: S. Lewis-Simpson (ed.) Vínland Revisited: the Norse World at the Turn of the First Millennium. Selected Papers from the Viking Millennium International Symposium, 15-24 September 2000, Newfoundland and Labrador, St John's: 165-180. - 19 - 8/01/2007 h:01 a1/p1 J.F. Wilson, D.A. Weiss, M. Richards, M.G. Thomas, N. Bradman, and D.B. Goldstein (2001) Genetic evidence for different male and female roles during cultural transitions in the British Isles, Proceedings of the National Academy of Sciences 98: 5078-5083. Abbreviations EurIsles: EurIsles – European Islands System of Links and Exchange (www.eurisles.com). SIC 2003: Shetland Islands Council (2003) Shetland in Statistics, Lerwick (http://www.shetland.gov.uk/council/documents/18170-Shet-in-Statistics.pdf). StatGreen: Greenland Statistics (www.statgreen.gl/english/). UNEP: United Nations Environment Programme Islands website (http://islands.unep.ch). UNSD: United Nations Statistics Division, Environmental Indicators, Land Use (http://unstats.un.org/unsd/environment/agriculturalland.htm). Figures and tables Fig. 1. Basic schematic depiction of migration (Lee 1966: 50 chart 1). area (a) IA = distance area suited for livestock tending x 100 (b) IA = LSV x total distance Fig. 2. Formulae used for the calculation of the index of attractiveness (IA): (a) standard index of attractiveness; (b) index of attractiveness used in table 4. Total area Livestock (km²) Reference Livestock (km²) Reference Rank Orkney 956 EurIsles 100% 956 See text 5 Shetland 1468 SIC 2003 100% 1468 See text 3 Faroes 1399 UNEP 100% 1399 See text 4 Eysteinsson and Iceland 101826 UNEP 27% 27493 Blöndal 2003 1 Greenland (ice-free) 410449 StatGreen 0,6% 2350 UNSD 2 Table 1. Biogeographical rankings according to area suitable for livestock farming. Distance from/to (km) Norway Orkney 420 Orkney Shetland 320 120 Shetland Faroes 600 320 310 Faroe Iceland 970 820 800 460 Iceland Greenland / / / / 350 - 20 - 8/01/2007 h:01 a1/p1 Table 2. Approximate direct distances in km between Norway and the North Atlantic islands (measured with the Google Earth measuring tool). From To Stepping Stone LSV (km) Total Distance (km) Norway Orkney None 420 420 Norway Shetland None 320 320 Norway Faroes Shetland 320 630 Norway Iceland Shetland + Faroes 460 1090 Iceland Greenland None 350 350 Table 3. Most likely sailing routes of the first colonists between Norway and the North Atlantic islands. Livestock (km²) LSV (km) Total Distance (km) IA Ranking Orkney 956 420 420 0,54 5 Shetland 1468 320 320 1,43 3 Faroes 1399 320 630 0,69 4 Iceland 27493 460 1090 5,48 1 Greenland 2350 350 350 1,92 2 Table 4. Index of attractiveness (IA) and biogeographical ranking of the North Atlantic islands. (Attempted) Exploration Discovery Colonization stage Ranking Orkney < 800 AD < 850 AD >? 50 years 4 Shetland < 800 AD < 850 AD >? 50 years 4 Faroes 825 AD* 850 AD 25 years 3 Iceland 860 AD* 865 AD* 5 years 1 Greenland 920 AD* 985 AD* 65 years 6 Newfoundland 985 AD* 1000 AD* 15 years 2 Table 5. Chronological ranking based on duration of the exploration stage. Ranking Biogeographical Chronological Orkney 5 4 Shetland 3 4 Faroes 4 3 Iceland 1 1 Greenland 2 6 Newfoundland / 2 Table 6. Confronting the biogeographical and chronological rankings. - 21 - 8/01/2007 h:01 a1/p1