On Some Pragmatic and Symbolic Features of Island Representation in Old Norse Literature

Book 7 of the classical masterpiece The Odyssey contains a scene where the hero Odysseus (Ulysses) describes how he was shipwrecked and washed ashore on Calypso’s island, Ogygia. He thus relates: “[...] There is an isle, Ogygia, which lies far off in the sea. [245] Therein dwells the fair-tressed daughter of Atlas, guileful Calypso, a dread goddess, and with her no one either of gods or mortals hath aught to do; but me in my wretchedness did fate bring to her hearth alone, for Zeus had smitten my swift ship with his bright thunderbolt, [250] and had shattered it in the midst of the wine-dark sea.”1 Calypso kept Odysseus with her for seven years and even promised him immortality if he would stay on the island – an offer refused by Odysseus, who longed to return to his family. At last, with Zeus’ intervention, Calypso had to let Odysseus leave the island – but his adventures were far from being over. The reason for using the quote above as an introduction to the present paper lies in the fact that it illustrates fittingly the central role that islands have played in the human mind and imagination for a long time. The sequence of events concerning the numerous islands that Odysseus reached during his voyage is perhaps one of the most well-known cases; already there and then one can trace characteristic motifs developed in western thought in connection with islands. Another popular example is the medieval Navi ga tio Brendani, a story about the legendary voyage of Brendan of Clonfert, a sixthcentury Irish monk who is said to have travelled from one island to another in search of the Isle of the Blessed (the Promised Land of the Saints).2

During different times and within varying cultural contexts the themes and approaches can naturally vary a lot. But at the same time, there is ample reason to claim that there has existed and still exists a universal fascination with islands. This is what we could call the phenomenon of 'islomania' characterised as "[…] a central feature of Western culture, a core idea that has been a driving force from ancient times to the present" (Gillis 2004: 1). According to Gillis, such an attraction does not simply refer to being interested in islands as particular features of landscape, but also to a deeper epistemological experience: "Dividing the world into discrete things, islanding it as a means of understanding, is a peculiarly Western way of navigating a world that seems otherwise without shape and direction" (op. cit.: 2).
The main focus of this paper concerns the discourse of islands on the basis of medieval Old Norse-Icelandic narrative tradition -one chapter in the history of island representation. According to our view, the imagery of islands in Old Norse sources is simultaneously reflective of certain contemporary European ideas, as well as of specific Nordic experiences that also bear witness to cultural-historical awareness.3 The emphasis in this paper is twofold: for one, we shall concentrate upon the representation of particular islands and events. Such a perspective serves at the same time to illuminate the above-mentioned Nordic dimension of island experiences. Secondly, we shall discuss the manner in which certain more general and symbolic features can be attached to the concept of island as mediated by the sources. This latter viewpoint can parallelly be applied to illustrate some facets of a broader medieval tradition on islands.

Theoretical considerations
In the introduction the source material was defined as medieval Old Norse-Icelandic narrative tradition. Needless to say, this paper will discuss only a limited fraction of the overall material. We have chosen to use particular forms of saga literature as the point of departure -more precisely, the main emphasis is laid upon relevant examples deriving from the sagas of Icelanders (Íslendingasögur) and the kings' sagas (konungasögur) -as is well known, the former focus upon the activities (conflicts) of significant Icelandic families mainly in the 10th and early 11th centuries, whereas the latter relate the history of primarily Norwegian kings through different periods of time.4 With the sagas themselves being written down some time during the 12th-14th centuries, they treat the subject matter from a retrospective and predominantly realistic point of view, making the claim of being "historical sagas about the past," to quote Meulengracht Sørensen (1993: 98).5 In addition to discussing the island imagery in sagas, there will be a few parallels drawn to the motifs occurring in Old Norse poetry -meaning first and foremost the skaldic praise poetry of the Viking Age and the early Middle Ages. It is logical to include corresponding examples due to the very fact that they are mediated as quotations in the same sagas. It is indeed thanks to the eagerness of medieval saga writers to illustrate and document their statements in terms of (potentially authentic) skaldic stanzas that this category of Old Norse poetry has been preserved.6 But besides skaldic poetry, we shall also briefly refer to the island theme in the framework of the mythological tradition of eddic 4 The temporal frames of individual kings' sagas are summarised e. g. by Jackson (1993: 10-15). A considerable part of the kings' saga tradition centres around the two missionary Norwegian kings, i. e. Óláfr Tryggvason (ruled ca. 995-1000) and Óláfr Haraldsson (ruled ca. 1015-1030). 5 Another matter is whether sagas can be analysed as historical sources as well.
During the past few decades scholars have started to treat sagas as potential sources of cultural history and the history of mentality. For a discussion of different trends in recent saga scholarship, see e. g. Lönnroth (1993). 6 The authenticity and credibility of preserved skaldic poetry is another widely discussed matter. One common view is summarised by Frank (1985: 173): "Today almost none of the verse in the family sagas is considered secure; poetry in the kings' sagas still commands credence, for it has not yet seemed likely that these verses are fabrications, falsely attributed to the early skalds." However, the debates around individual poems continue. Concerning the source value of skaldic poetry and the criteria for tracing genuine stanzas, see also Bjarni Einarsson (1974);Vésteinn Ólasson (1987);Jesch (1993;2001). poetry as this was (still) known in medieval times.7 When consulting different types of sagas as well as poetic texts, it is possible to illustrate the application of particular themes across various genres of Old Norse literature and in this manner demonstrate their common character. In further studies the scope of the source material can be enwidened to include other forms of saga literature and medieval prose tradition.
The main arenas of action in Íslendingasögur and konungasögur are Iceland and Scandinavia, respectively. The latter area also lies in the focus of much of the skaldic praise poetry, whereas with regard to eddic poetry we cannot determine the precise setting in the same manner.8 Besides the main setting, sagas also include references to various other travel destinations -travel is in itself a popular motif in many a saga narrative; or as it has been put, the sagas "[…] are full of movement and experience, at home and abroad" ( Jesch 2005: 134). The image of a travelling poet, chieftain, king or simply a fame-and fortune-seeking young man is equally well recorded in Íslendingasögur and konungasögur; both types of sources cast light upon movements on a more regular scale, such as smaller raiding and trading enterprises, as well as upon larger campaigns and expeditions. Similarly, much of the skaldic poetry contains elements of travelogue in terms of listing the sites where the honoured kings and chieftains headed to. As expected, the content is then poetically modified; the act of travelling interests the skald in the framework of dramatic events and is combined with expressive battle imagery.
In this current context we shall examine islands as belonging within the scheme of travelling. This allows us to draw attention to several characteristic features of the overall island representation. In order to limit the scope of study, we have chosen to concentrate mainly upon two geographical settings. For one, we look at the islands in the Baltic Sea, in the region of the Kattegat and the Danish straits -an active arena for communication. The second setting is made up by islands along the western coast of Norway that are depicted to host frequent traffic in the context of western travels along the North and the Norwegian Seas.9 Both of the above-mentioned territories demonstrate clear communicative significance as attested to by the sagas.
A final consideration concerns the definition of an island in this study. Determining what exactly an island is can be a tricky linguisticsemantic, cultural-historical as well as geographical matter. On the one hand, one could analyse the application of particular linguistic elements in Old Norse place names and look at various descriptive references provided by the sources.10 On the other hand, it would be necessary to take into consideration different contextual factors and among other things to discuss the relationship between island(s) and mainland(s). Certain puzzlement concerning the distinction between islands and mainlands is visible from the sources as well. In the 13th-century Norwegian prose work known as Konungs skuggsjá (The King's Mirror) -composed as a dialogue between a father and a son -a question is posed as to whether Greenland should be considered a mainland or an island: "Svá forvitnar mik ok þat, hvárt þér tlit at þat sé meginland eða eyland" (Keyser et al 1848: 42).11 The father then explains that the size of Greenland is unknown, but it is taken to be a mainland and that it is connected to some other mainland; the proof is found in the fact that Greenland has such animals that do not usually live on islands. That one had to get a sense of a territory in order to determine whether this was an island also shines through in a short passage in Groenlendinga saga (ch. 2) describing the voyage of Bjarni Herjólfsson. Bjarni and his men intended to sail to Greenland, but due to unfavourable sailing conditions came across several unknown lands. Approaching a third unknown territory they: "[…] halda með landinu fram ok sá, at þat var eyland" (Matthías Þórðarson and Einar Ólafur Sveinsson 1935: 247).12 A few general concepts to keep in mind in the meantime are the following. In Old Norse cosmography (and similarly in ancient geography), we find the notion of the inhabited world being an island, surrounded by the chaotic Ocean, in Old Norse known as Úthaf (see e. g. Hastrup 1990: 28-29;Gillis 2004: 12-13). Considering the model that sees the whole world as an island, it seems obvious that the size (and the status) of the landmass is not necessarily an apparent feature to pursue when defining (is)lands. Also, an island does not even have to be surrounded by water; according to Gillis (2004: 17), in medieval times any strange and distant place could be viewed as an island.13 Here we shall follow a somewhat simplified practice and still define islands first and foremost as waterbound insular communities, emerging as such on the basis of the narrative context itself. This approach is motivated by the actual perspective of the sources; as is for example said in the 13th-century work known as Snorra Edda or The Younger Edda (according to the manuscript Codex Upsaliensis, DG 11): "[…] en ey heitir þat land sem sior eða vatn fellr vm hverfis" ( Jón Sigurðsson 1852: 366).14

Islands as significant sites
The treatment of islands in Íslendingasögur and konungasögur as well as in skaldic poetry in many ways presents them as useful sites for localising various events. Furthermore, they emerge as important cognitive landmarks from the maritime perspective, bearing witness to the sailing experiences of the Northmen. On the whole, we can say that the role of islands in the narrative is very much connected to the typical activities of a travelling hero. One distinctive facet of island imagery in the sagas and in skaldic poetry is that they provide a suitable setting for battles, campaigns and raids. The fighting may then occur either directly on the island or in the waters around it; various features of the landscape may be illuminated, showing for example how islands could be used to organise stakeouts, etc. Corresponding motifs are well-recorded in konungasögur, but also meet us in certain Íslendingasögur.
At one extreme, such island confrontations include big battles between Scandinavian kings; at the other, private duels between two opponents. In fact, the concept of duelling creates an association to islands also in terms of its name, i. e. hólmganga, which contains a reference to small islands (holmr / hólmr). As we learn from the sagas, an alternative to the island setting is to carry out a duel on top of a small hillock; in the meantime, even such a site appears as a kind of island in relation to the surrounding landscape. But the sagas do also refer to actual island duels -and those may involve fighting a vicious warrior or even a supernatural creature. 15 We shall illustrate the motif of islands as key (battle) sites by taking a look at the island of Hlésey (Laesø) in the middle of the Kattegat. In the saga on Magnús blindi and Haraldr gilli in the 13th-century Norwegian kings' saga compilation Heimskringla, we hear about a battle Haraldr held by that island (see ch. 12 of Magnúss saga blinda ok Haralds gilla). The same event is referred to by the somewhat earlier kings' saga compilations Morkinskinna and Fagrskinna. In all three cases the saga statements are illustrated by a skaldic quote assigned to a 12th-century poet Einarr Skúlason; here the whole stanza is given according to the skaldic poetry edition: "Ǫ ttuð sókn við sléttan, / serkrjóðr Hǫars, merki, / harðr, þars hregg of virðum, / Hléseyjar þrǫm, blésu; / hús brann up, en eisur / ófatt, séa knátti, / malmr sǫng, en hlóð hilmir / hr kǫst, við ský gn fa" (Finnur Jónsson 1912-1915.16 Before mentioning the battle at Laesø, the sagas point out another island and again an illustrative skaldic quote is inserted into the narrative. This island is identified as Hveðn (Ven), a small island in the strait of Øresund, between Sjaelland and Skåne. Fagrskinna also specifies here that Haraldr was fighting against some Vikings.
In this context islands are thus identified as fitting sites for holding victorious battles. However, island encounters do not always prove to be successful for the saga character; the narratives also highlight potential dangers connected to such places. In ch. 94 of Óláfs saga helga in Heimskringla, we hear about a certain Gauti Tófason; his story is told as news to the Swedish king by a wise man called Emundr from Skara.17 One time Gauti Tófason sailed along the Göta River and when he reached Eikreyjar (referring to the northern Göteborg archipelago where we also find the island of Öckerö) he noticed five Danish trade ships. Gauti conquered four of them together with his men, and then started chasing the last one with one of his ships, but lost sight of it. Due to a heavy storm he lost his ship and all the men on board -this happened by the island of Laesø. Meanwhile, his remaining companions who were still waiting for him by Eikreyjar themselves got attacked by more Danes, and now the fortune turned, because they all got killed. Besides its informative value this little tale carries allegoric significance, since it demonstrates how one loses everything as a result of having got too greedy. In this manner, Emundr also prepares the ground for bringing up a more important case of complaint in front of the king. This latter example referred to Laesø in the context of stormy weather leading to a shipwreck. Islands can thus figure in the scheme of events that highlight the hardships caused by weather and problematic sailing conditions -in a situation like that, battling becomes much more challenging. Laesø is also mentioned in ch. 35 of Haralds saga Sigurðarsonar in Heimskringla. King Haraldr, who has been raiding around in Denmark, is sailing northwards in the Kattegat when contrary winds force him to take lee by Laesø. The same scene is brought up by Fagrskinna (see ch. 55) and Morkinskinna (p. 166). In the meantime, the latter two accounts claim that Haraldr had to find shelter by a different Danish island -namely Sámsey (Samsø). In all three cases there is also mention of a heavy fog occurring on the island that forms a contrast to the sun reflecting from the ships out on the sea -those being the approaching ships of Haraldr's opponent, the Danish king Sveinn Úlfsson. As explained by Finlay (2004: 206), the reference to Laesø in Heimskringla can be considered "more probable in relation to Limafjǫrðr" (i. e. the sound of Limfjord in Jutland), also forming part of the setting. On the other hand, Samsø gets mentioned as an important locality earlier in Morkinskinna in connection with King Magnús góði; the island is then characterised as the usual anchorage for the king: "[…] þeir kuomu uid Samsey og lagu skipunum þar j einre hofn sem jafnan hafde legit Magnus konungr fyrr" (Finnur Jónsson 1932b: 146).18 It is not always necessary to identify the location precisely; more important is to show that the site as such fulfils its particular narrative purpose. On certain occasions, it can also be expected that specific contextual clues would make it clear which place one had in mind. Among famous battle sites, the island of Svǫlðr (i. e. Svolder) is given as the setting for the final battle of Óláfr Tryggvason.19 As for example stated in Fagrskinna (ch. 24), the enemies of the king had gathered their forces there: "Við einn hólma fyrir Vinðlandi váru saman komnir margir stórir hǫfðingjar. Þessi hólmi heitir Svǫlðr" (Bjarni Einarsson 1985: 147).20 In the preserved skaldic poetry relating of that battle, the name of the island is not necessarily specified. Erfidrápa Óláfs Tryggvasonar by Hallfreðr Óttarsson vandraeðaskáld, for example, localises the battle in the south of the sea; the location is further described as the broad sound of an island (á víðu Holms sundi, see stanza 17). The southern direction is also emphasised in Eiríksflokkr by Haldórr ókristni; and there as well a

"[…] they came to Sámsey (Samsø) and anchored in the harbor where King
Magnús had always anchored off that island in earlier days" (Andersson and Gade 2000: 185). 19 The exact location of Svolder is unknown; it has for example been suggested that it is an island somewhere in the southern Baltic close to the German coast. See the short overview provided by Andersson in his translation of The Saga of Olaf Tryggvason by Oddr Snorrason (2003: 147); cf. also Finlay (2004: 116). 20 "By an island off the coast of Vinðland there were gathered many important chieftains. This island is called Svǫlðr" (Finlay 2004: 116). See also ch. 99 of Óláfs saga Tryggvasonar in Heimskringla.
reference is provided to the island in terms of applying the word holmr (see stanza 3). Perhaps in these cases it is the southern setting itself that makes the locality clear, so that no further identification is required.21 On the other hand, sometimes the sagas may find it important to identify even small and uninhabited islands and provide descriptive comments as to their particular features. One such example concerns Brenneyjar (in modern Swedish Brännöarna) -another archipelago on the western coast of Sweden. These small islands are mentioned on several occasions; and in Bjarnar saga (ch. 7) it is explained: "[…] þat eru margar eyjar ok váru þá lítt byggðar. Þar váru í launvágar, ok var þar jafnan herskátt af víkingum; skógr var þar ok nǫkkurr á eyjunni" (Guðni Jónsson and Sigurður Nordal 1938: 127).22 Similarly in ch. 48 of Egils saga Skalla-Grímssonar, we hear that Vikings used to wait for passing trade ships in the region of Brenneyjar.
The examples above have illustrated certain islands as the focal points in the southern Baltic. Other islands in the same region (but also in other maritime landscapes) fulfil similar functions. In the case of the bigger Danish islands, such as Sjaelland, Fyn, and Falster, it is obvious that besides being common battle sites they also belong with the picture of extending one's territorial dominion -much traffic is taking place in between these and other islands as a result of the political confrontations between various Danish and Norwegian rulers. But in addition to that, the depiction of islands in the Baltic Sea also expresses strategic navigational purposes. The islands form part of common sailing routes, and the knowledge connected to them has pragmatical significance.23 On the one hand, islands may appear as navigational aids, suitable outposts and anchorages; on the other hand, it is important to stay aware of the challenges that sailing around certain islands could pose. Such are the very practical aspects of island representation that can be traced on the basis of both narrative and directly geographical sources.
In the following, we shall concentrate upon a more symbolic facet of island representation. Earlier we referred to such island combats where one has to fight against some remarkable or even supernatural creature. A quote from eddic poetry that refers to the island of Laesø from a similar perspective can provide a fitting transition from the practical to the symbolic. Thus, in the poem Hárbarðsljóð, stanza 37, we hear about the god Þórr fighting against some berserk women on Laesøcombining a clear identification of a well-known maritime site with mythological motifs. This is what Þórr says about his island experience: "Brúðir bersęrkja / barðak í Hlésęyju, / þ r hǫfðu verst unnit, / vélta þjóð alla" (Finnur Jónsson 1932a: 87).24

Island symbolics
This brings us to the symbolic level of island representation, which in many ways accords with a broader medieval tradition on islands. Islands thus get connected with peculiar events and appear as mysterious and miraculous sites -both in the positive and negative sense. To start with the latter, in sagas we for example meet the motif of islands as sites for outlaws and criminals. Due to its relative isolation, an island can provide a perfect hiding place and / or prevent contact. Similar perceptions shine through in an island reference occurring among eddic poems, namely in Vǫlundarkviða (see Finnur Jónsson 1932a: 125-126). A small fictional island called S varstǫð (meaning 'sea-harbour') figures as the place where the smith Vǫlundr is kept in captivity and has to work for the king. No one else is supposed to have contact with him; this does not really work out, and later in the poem we hear about some horrific incidents taking place on the island.
Another motif concerns the somewhat strange habits of people living on islands, as well as the fact that islands may provide a home base for weird and dangerous creatures. With regard to the former, Old Norse even had a special term coined for marking islanders -they could be called eyjar skeggjar, i. e. 'island beards' (see Cleasby et al 1957: 134). Concerning the latter aspect, we already referred to certain motifs in sagas (for example, Grettis saga Ásmundarsonar) that speak about encounters with island monsters.
In accordance with common medieval perceptions, an island itself may appear as a monster or a gigantic creature of some kind. The abovementioned Navigatio Brendani includes a scene concerning the so-called Fish Island ( Jasconius). Brendan and his followers camp on the back of the giant fish, believing it to be an island; they even celebrate Easter there. Once they light a fire on its back, the monster awakes and starts moving and the men have to flee.
Parallels to that particular story and / or other similar motifs can perhaps be found in a little humouristic episode related in ch. 3 of Knýtlinga saga and ch. 33 of Óláfs saga Tryggvasonar in Heimskringla. In this connection, we hear about problems the Danish king Haraldr Gormsson was having with Icelanders who were producing mocking verses of him. One of the king's sorcerers is sent on a magic ride to Iceland in the form of a whale; remarkably enough, he later informs the king that there live all kinds of monstrous creatures on that land, and that it also lies way too far to be reached by ship. This little comment provides a kind of side-step regarding the image of Iceland, in supposedly representing the island's role in the eyes of those from the outside, and fitting well with the medieval practice of depicting remote and mysterious places.
But islands could also be given religious significance; according to another medieval concept islands emerge as holy and sacred sites (see Gillis 2004: 26-39). In Old Norse sources islands may be associated with both heathen and Christian practices; on more specific occasions there are even magic activities taking place there. Looking at an example from eddic poetry, in stanza 24 of Lokasenna Loki accuses Óðinn of having engaged in disgraceful magic practices (seiðr) on the island of Samsø: "Ęn þik síða / kóðu Sámsęyju í, / ok drapt á vett sem vǫlur, / vitka líki / fórt verþjóð yfir, / ok hugðak þat args aðal" (Finnur Jónsson 1932a: 105).25 Again, it is interesting to observe how the very real island is connected to certain fantastic events in the framework of eddic poetry.
In this current context we shall in the meantime concentrate upon the motif of islands as sites for religious transformations resulting from the act of Christianisation. The islands that are in focus are those located along the western coast of Norway connected to the missionary activities of Óláfr Tryggvason and Óláfr Haraldsson. We shall use the former as our point of departure and look at a few scenes occurring in a separate saga on Óláfs saga Tryggvasonar by Oddr Snorrason (from the end of the 12th century), as well as in the kings' saga compilations Fagrskinna and Heimskringla.26 Óláfr is said to have sailed from the Orkney Islands to Norway after having baptised the Orkney jarl. In Norway he lands on the island of Moster (in Sunnhordland); during the night he is visited by a saint who says the king should call people to give up their old customs and accept the new religion. We further hear about Óláfr organising an assembly on the island and speaking to the people while standing on a high rock. Three men attempt to come with arguments that would oppose Óláfr, but neither among them is able to deliver his speech: "Nu firir þui at þeir voro sigraðir með sua miclum krapti. Þa tocu margir við tru. Oc firir letu forna villu oc fylgþu allir konungs boði" (Finnur Jónsson 1932c: 96).27 In the following, the saga relates about the island of Selja (in Sogn og Fjordane), which emerges as a sacred site. Several people are said to have experienced special light and sweet fragrance coming from the island, and they relate these stories to the king. Óláfr then heads to the island together with his bishop, and they find there a place with bones, i. e. holy relics. The king understands the religious significance of the island and thus has a church built there: "Oc at bỏn byskups oc konungs raði. var þar kirkia ger. oc helguð þessum guðs monnum er þar voro" (Finnur Jónsson 1932c: 100).28 The saga also provides a comment with regard to the island of Kinn: "Slict hit sama verþa oc morg tocn iannarri eyio er KiN heitir. ero þar oc helgir domar þessarrar sueitar. sem iSelio. Oc ueitir guð firir huarratueggio sakir margar iartegnir firir milldi sina oc miscunn" (ibid.).29 Regarding the religious community of Selja, the saga includes a story concerning the tradition on Saint Sunnifa. According to the legend, she was an Irish princess who abandoned her home to escape marrying a heathen man. By the good will of God, she and her people came to the island of Selja. Selja also became the death place of Sunnifa and her followers -the caves on the island collapsed over them, saving them from falling into the hands of local heathens who thought they were causing trouble on the island. The saga also remarks that there are many miracles connected with the island and provides the names of the churches that have been built there.30 As a parallel to the discussed episodes in Óláfs saga Tryggvasonar by Oddr Snorrason, we can mention that ch. 47 of Óláfs saga Tryggvasonar in Heimskringla contains a short reference to the island of Moster. The island is similarly identified as the site where Óláfr Tryggvason first came ashore. According to Heimskringla, a mass was sung on the island, and later a church was built there.
Heimskringla does not relate of Selja in connection with Óláfr Tryggvason. In the meantime, Selja is mentioned in ch. 29 of Óláfs saga helga -with certain parallels noticeable between this reference to the activities of Óláfr Haraldsson and the one that relates of Óláfr Tryggvason's arrival on Moster. Both islands function as outposts for the kings arriving in Norway. In the case of Óláfs saga helga, it is told 28 "At the request on the bishop and the command of the king a church was built there and dedicated to the men of God who were there" (Andersson 2003: 77). 29 "There are also many signs on another island called Kinn. There are also holy relics of the same community as on Selja. In honor of both, God performs many miracles in his graciousness and mercy" (Andersson 2003: 77). 30 A detailed account on the tradition concerning Sunnifa and the actual finds made on the island of Selja is given in Rindal (ed., 1997).
that during the return trip to Norway Ólafr and his men experienced trouble on the sea, but thanks to the good luck of the king everything went well. They landed on the island, which in Heimskringla is given the name S la, meaning luck and happiness. This is naturally taken as a good sign by the king. Fagrskinna contains a similar scene in connection with Óláfr Haraldsson (see ch. 28): "Þeir kómu at hafi útan at Staði ok þar á land, sem ey ein lítil er ok heitir S la. Þá m lti Óláfr ok lét þá tímadag hafa land tekit ok talði þat gott mark, at þeir váru komnir í S lu" (Bjarni Einarsson 1985: 170).31 In both Heimskringla and Fagrskinna, it is further related how the king stumbled on land, an unlucky accident.
In the meantime, his companion tried to save the situation by stating that in this manner the king established his power over Norway.
The symbolics connected to the island of Selja as it is represented in saga narratives, as well as other sources that relate of Sunnifa, demonstrates that the site was considered important for several reasons. The explicit saga motifs show Selja as a holy site, with clear religious significance. From a comparative perspective, it has been pointed out that one can find elucidative parallels between Selja and certain holy sites along the British coasts. As argued by Crawford (1997: 178), it must have been necessary for the kings to find a suitable locality for establishing a religious cult; and according to the Celtic tradition, an island off the coast was especially well suited for such purposes. In this way, we can also notice missionary and political motivation behind the scenes that are reflected in the sagas. Also, the emphasis upon Selja and other similar islands is not accidental. As has been pointed out for example by Hommedal (1997: 63-65), the location of Selja is indeed strategic, and the island made up a perfect natural harbour for people travelling past the Stad peninsula. The latter promontory is known because of complicated weather conditions that can occur in its neighbourhood. Thus, for travellers Selja must have made up an ideal anchorage when they had to wait out harsh weather, for 31 "They came in from the sea at Staðr, and came to land where there is a little island called Saela. Then Óláfr spoke, saying that the day of their reaching land was a lucky one, and he reckoned it a good sign that they had come to Saela (Happiness)" (Finlay 2004: 137). As commented by Finlay (ibid.), "S la has been considered a variant of Selja, adopted for the sake of the pun".
example.32 In this way the circle is complete, since we have returned from symbolic and religious features to the visibly pragmatical aspects of island representation.

Conclusions
Island representation in Old Norse narrative sources is based upon insights from the practical activities and experiences of the Northmen as well as the more general medieval concepts concerning communication with islands. In this current paper we have concentrated upon a few limited aspects of such island imagery.
Islands -both real and imaginary, named and unnamed -appear in different types of narrative contexts and fulfil a variety of functions. On the one hand, islands may emerge as significant (battle) sites that are further characterised by their strategic positioning on the maritime landscape; on the other hand, it has been shown that there are special religious and mythological motifs attached to them, which illuminate their peculiarity.
The examples discussed above have at the same time demonstrated that one dimension does not have to exclude the other -we can rather witness a combination of various types of features across the sources. This concerns both the imagery of particular islands (take for example the Norwegian island of Selja), as well as the broader understandings on islands as spatial and cognitive figures. In this we can witness a mixture of various narrative depiction techniques, as well as the blending of actual cultural traditions.
All in all, it is obvious that the Old Norse sources in their own way bring out the many dynamic qualities of island representation in the human imagination. As such, there are various ways in which the mediated images can contribute to our ongoing fascination with islands.