CROSSES ON RUNESTONES Functions and Interpretations

87 Crosses on Swedish rune-stones have been studied on numerous occasions, mostly in isolation from other features of the monument. This article examines the use of rune-stone crosses with an emphasis upon their varying functions in the total composition of runic monuments. The analysis that combines the level of visual composition with textual elements reveals different strategies in the display of crosses. Besides functioning as externalized Christian markers, crosses could be made to serve various internal (i.e. inscription-based) stylistic, decorative and practical purposes. The role of the cross could be modified according to particular contexts of usage.

Cross ornamentation on Swedish rune-stones has been the subject of several studies of archaeological, art-historical, cultural-historical and runological orientation. The focus has lain on the classification of cross shapes (e.g. Liljegren 1832:141ff, plate 5; Plutzar 1924, plate 19;Gardell 1945:62ff, 73ff, 82f;Wideen 1955:147ff), combined with discussions of their chronology, provenance and sources of influence. In a study of Upplandic rune-stones, Claiborne W. Thompson identified six basic structural types of rune-stone crosses (Thompson 1975:30f). His classification provided a common frame of reference until a more flexible way of analysing rune-stone crosses in terms of different variables was presented by Linn Lager (2002).
In the context of using rune-stones as sources for the introduction and spread of Christianity, crosses have been characterized as explicit

CROSSES ON RUNE-STONES Functions and Interpretations
Kristel Zilmer markers of the Christian faith. As is well known, the inclusion of cross ornamentation is the most frequent Christian feature on runic monuments; based upon Lager's corpus (2002:73ff, 95f), 58% of rune-stones in Sweden have crosses. According to one theory, cross-marked stones could serve to consecrate burial grounds (Gräslund 1987:258). The analysis of rune-stone crosses in combination with other elements of ornamentation has shown that they were adapted within the unique setting of "Scandinavian visual language" (Lager 2002:193ff;Lager 2004:150f). In this way, the cross was integrated into the Scandinavian context.
It has been customary to analyse rune-stone crosses in isolation from the rest of the monument due to their character as distinct visual symbols. Studies that discuss the position of the cross in relation to the design of the runic carving have broadened this approach (Plutzar 1924:33ff;cf. Lager 2002). The discussion about the function(s) of rune-stone crosses can be brought further from a perspective that combines the evident visual level with the composition of the monument and the verbal level of the runic text.
Various strategies of using the cross in the composition of runestones can be observed in the material analysed here. The analysis will cast light upon some functional attitudes towards cross imagery on the part of the people who made the decision to employ this ornament and/ or supervised the execution of the carving on the stone in question. On this basis it can further be discussed whether crosses on rune-stones automatically functioned as externalized visual markers of one's religious adherence or whether we should also take into consideration various internal (i.e. inscription-based) stylistic, decorative and practical motives. One issue concerns the potential gestural function of the cross. In studies that explore the interplay between imagery and text, it has been suggested that the placing of crosses on rune-stones created meaningful symbolic links with runic texts (Andrén 2000:18ff;Lund 2005:121ff). I will examine certain claims made about connections between cross imagery and elements of runic texts. Special attention will be devoted to the use of runes inside the cross.
This article deals mainly with rune-stones from central Sweden, from the provinces of Södermanland and Uppland, since this area offers the most extensive study material on the use of crosses. Additional examples are given from other runic regions of Sweden in order to illustrate relevant occurrences elsewhere. The temporal frames coincide with the main era of raised rune-stones, from the end of the 10th century until the beginning of the 12th century, with occasional mention of earlier or later material. In terms of broadening the discussion of rune-stone Crosses on Rune-Stones crosses, some comparisons will be drawn regarding the use of crosses on contemporaneous and younger grave monuments, respectively.

CROSSES IN THE VISUAL COMPOSITION OF RUNIC MONUMENTS
The first part of the article deals with varying strategies that are used in the display of crosses, partly in accordance with observations made in previous studies (e.g. Lager 2002). The qualitative survey serves to exemplify and underline the fact that the role of the cross in the composition of runic monuments varied -on the one hand, crosses could figure as independent ornaments, on the other hand, there existed various ways of making the cross an integrated part of the carving. Additional design-related features have to do with the size and number of crosses as well as the use of crosses on different sides of one runic monument and on paired monuments. All such factors could influence the significance attached to the cross as an element of ornamentation and also modify its role as a symbolic marker of Christianity.

The cross as an independent ornament and symbol
The dominant visual role of the cross is evident from its customary position in the central or upper part of the stone. The cross forms a separate visual entity that is not directly connected with the runic inscription or any ornamental elements. As an eye-catching emblem, the cross must have been an easily recognizable visual symbol, and it can be considered an externalized marker of Christianity. This type of layout is well established on rune-stones from Södermanland and Uppland as well as from other parts of Sweden, figure 1. It is also present on contemporaneous and younger runic grave monuments, though possibly modified according to the varying material features and the shape of the monument. For instance, on horizontal flat slabs the cross can be positioned on top of the monument, whereas the runic text runs along the narrow edges of the slab. 1 Designs similar to those of raised rune-stones can appear on upright standing runic gable slabs of various sizes (forming parts of different types of grave monuments, including the so-called Eskilstuna cists); the cross can stand in the centre of the stone, framed by the runic inscription. 2 On the other hand, gable slabs can be reserved for the Crosses on Rune-Stones cross (or other ornamentation), whereas runic inscriptions are placed along the sides of the richly ornamented lid slabs; the latter strategy is indeed more common (cf. Lundberg 1997;Ljung 2010; on the iconography of Eskilstuna cists and other types of Christian grave monuments, see also Staecker 2010:243ff).
Another recurring feature that highlights the visual status of the cross is its placement at the very top of the stone. Crosses in this position may stand on their own, lean slightly upon the band of runes or be lifted above it with some help from the text band/runic serpent. 3 In contrast, it is not common to place the cross in a position that would diminish its visual prominence. Separate crosses that appear close to the base of the monument (and below the runic inscription) can be smaller crossmarks. 4 These resemble cross-shaped signs that could be used as (word) dividers/space markers within the runic text. The considerable variation that occurs in the use of cross-shaped signs makes it impossible to prove that they fulfilled extended symbolic functions in rune-stone inscriptions. They may, however, mark the beginning of the inscription, or some part of it. 5 This technique, which speaks of certain literate conventions, is to a greater degree attested in medieval inscriptions, including those on runic grave monuments.
Returning to the independent significance and visual symbolism of the cross, it can also be made the focal point of the monument in terms of using other elements of the carving as gestural markers. An example of such intentional planning is possibly found in the Nora rock inscription (U 130), where the leg of the runic serpent guides the viewers' attention to the cross. 6 The described strategies of display show the role of the cross as a visual marker on rune-stones. In terms of being an externalized symbol of Christianity, its power of expression can be compared to that of free-standing stone crosses and cross slabs known, for example, from Norway and the British Isles. The Anglo-Saxon tradition relates that erected stone crosses could serve as sites where people could gather and pray, as described in the Hodoeporicon of Saint Willibald from before 786 (Talbot 1995:146;cf. DuBois 1999:148). Interpretations of the vari-3 Compare Sö 105, U 160 and U 1011 (both sides). 4 E.g. Sö 130 (small cross), Sö 367 (cross-mark), U 999 (two cross-marks). 5 A well-known example is the Karlevi stone (Öl 1, end of 10th century), with two cross-marks close to the base of the stone. One of them indicates the beginning of a memorial formula in prose, the other the start of a skaldic stanza. There is also a cross-sign in the Latin inscription (possibly a later addition) on a different side of the stone. 6 Cf. also U 328. ous functions of free-standing stone crosses that date from different periods and originate from different parts of northwest Europe vary. The stone crosses could mark graves and burial grounds as well as sites of worship (prior to churches); they could be used with the purpose of following Christian rites and teachings but also to express more practical considerations such as marking boundaries between properties (Nordeide 2009:174f). As such, the free-standing stone crosses could combine different traditions (Nordeide 2009:177).
Crosses at the top of raised rune-stones, or in other dominant positions, may also have functioned as visual aids, enhancing the significance of the commemorative monument in the surrounding landscape. The cross would at the same time highlight the meaning of the monument as a Christian site, and possibly as a place where one could say prayers for the deceased. As is known, commemorative texts on rune-stones are sometimes accompanied by prayers for the soul/spirit of the deceased; prayers are preserved on stones both with and without crosses. To name one example, the Långtora stone (U 804) has a central cross that stands on a horizontal line containing the runic prayer; no other text is found on the stone, but it probably belonged together with another stone that bore the commemorative text (U 839, see Stille 1999:192). The visual and verbal ways of communicating Christian concepts could have fulfilled complementary purposes in the reception of runic monuments. People who were not skilled in decoding runes could experience the visual meaning of the monument in terms of its cross ornamentation.

The cross as an integrated element
Besides forming an independent visual element on the stone, the cross can be integrated with the rest of the carving, that is, with different ornamental features and/or the runic band. The cross can be used to fill the inner space that is left vacant between the curving and/or crossing runic and ornamental bands, or more specifically, between the head(s) and tail(s) of the runic serpent(s). 7 The arms of the cross (one or several) may connect with the surrounding frame, either by touching against the band, cutting into it or merging with it (cf. group C in Lager 2002:70). One customary design ties the lower cross arm together with the text band or the body of the runic serpent close to the base of the monument. 8 The visual and perhaps even the symbolic significance of crosses that are in this manner integrated with the rest of the inscription vary. The deciding factors are the scheme of layout, the appearance of the cross, 7 E.g. Sö 109, Sö 126, U 654. 8 See e.g. the connected monuments U 135, U 136 and U 137. the size of the surface in which the cross is placed, and the patterns of interlace that may surround it.
A group of 37 rune-stones (the number is based upon the 2008 version of the Samnordisk runtextdatabas and includes some uncertain cases) is listed as belonging to the style group "cross-band stones" ("korsbandstenar", cf. Lindblad & Wirtén 1992). 9 The majority of cross-band stones originate from Södermanland, but they also occur in Uppland (9), Östergötland (3) and Småland (1). On these stones the lower cross arm, or alternatively an elongated staff that may be attached to the cross, is divided into two at the base of the monument and merges with the text band or the runic serpent. 10 The design makes internal use of the cross in the composition of the carving. This does not lessen the significance of the cross; often its proportions in relation to the rest of the carving, and indeed the whole raised monument, are impressive. 11 The cross still figures as an object that carries a meaning of its own. At the same time, it is of interest that this design makes it easy to extend the runic text into the cross (more about this below).
Specific designs connect with the styles of known rune carvers. On Upplandic rune-stones the cross can be placed inside the (upper) loop formed either by the curving body of the runic serpent or by additional ornamental bands. In this position, the arms of the cross can be attached to the surrounding runic or ornamental band. On the rune-stones signed by or attributed to the productive rune carver Öpir we often encounter the design of one or several curving loops; crosses on these stones are frequently placed inside the loops. 12 Another strategy is to depict crosses on rune-stones as partly or fully intertwined with zoomorphic motifs and/or interlace patterns. 13 The resulting visual impression may be complex, and it can be challenging to detect the cross on the monument -at least for someone with no expectations about its position. 14 The practice of displaying crosses in the context of typically Scandinavian zoomorphic ornamentation -well established on Upplandic rune-stones -has been interpreted as evidence 9 The study by Lindblad & Wirtén includes 55 rune-stones from Södermanland and Uppland, but not all of these would qualify as "cross-band stones" in the sense defined here. 10 Cf. e.g. Sö  of the cross gradually becoming a customized element of the Scandinavian visual language. As a result of such developments, the cross was no longer considered very important for marking the Christian identity of the rune-stone raisers (Lager 2002:195f, 222).
The broader impact of the custom of combining crosses with typically Scandinavian ornamental motifs is at the same time traceable in the context of runic grave monuments. Corresponding examples are, for instance, found among grave monuments from Östergötland. There, one can follow the tradition in its varying forms throughout the 11th century (on early Christian grave monuments in that area, see e.g. Ljung 2009b). The cross can be used as an integral element of the carving on different parts of grave monuments, such as gable slabs and lid slabs. At the same time, the cross does not form a compulsory element on grave monuments. Both early Christian grave monuments and later medieval grave monuments demonstrate that the inner space of the slab could also be filled with alternative ornamentation/imagery, or it could be left empty.
Rune-stones from Södermanland and Uppland, as well as from other parts of Sweden and Scandinavia, demonstrate the incorporation of crosses into varying visual contexts. The cross can be depicted together with boats, 15 birds (in some cases, possibly a rooster), 16 animals, 17 human figures, 18 and mask-like faces. 19 The interpretations of several such motifs have been debated (e.g. Lager 2002:187ff) and cannot be dealt with here. Such iconographical settings demonstrate a modified use of cross ornamentation, possibly arising from local and regional adaptations of Christian and/or other symbolism. The cross may, for example, be depicted together with motifs known from the mythological tradition of Sigurd the Dragon Slayer. 20 The placing of the cross within the context of various ornamental motifs presents it as an object of decoration that fills the inner space of the monument and works together with other imagery. Such designs do not emphasize the independent role of the cross, but rather focus upon internal links with the rest of the carving. This reveals a more flexible attitude towards the visual role and the symbolic meaning of the cross.

Crosses on Rune-Stones
The cross did not automatically have to stand in the position that would underline its function as a distinctive religious marker. It could be employed as an element of ornamentation, alongside other ornamental motifs. However, the decorative and stylistic purposes did not completely deplete the cross of its deeper meaning. What we witness here is rather an integration of different functions.

Single cross and multiple crosses
Special attention has to be devoted to rune-stones where the cross and the runic text are carved on different (opposite) sides of the monument. One such example is the Eggeby stone (U 69), on one side of which we find the memorial text and on the other a large cross. Various strategies of dividing verbal and visual elements between the two or three sides of the stone are used. 21 Besides the notable practice of using runes on one side and letting the cross stand on the other, there occur more complex schemes of division. On one side of the Landshammar stone (Sö 167), the runic serpent forms an arch along the monument. An image of a mask is fitted into the inner surface, together with a row of cipher runes. On the other side of the stone a separate cross is found -possibly placed there because the space on the runic side was reserved for other purposes. Such designs may have their own underlying logic; they do not seem to be completely accidental, but may reflect planned stages. When the main runic text and the cross appear on different sides of a stone, the reception of the monument must have been affected. People who approached the rune-stone from different directions could gain different first impressions of the monument. This could have been the case with the Eggeby and Landshammar stones.
In the case of paired monuments there occur further possibilities for dividing verbal elements and visual imagery. The strategies may have been based upon the choices of people who commissioned multiple monuments. Sometimes one of the stones bears the cross, but the other stone does not. The Harby stones (U 267 & U 268) have connected inscriptions -one inscription mediates the commemoration, the other displays the cross and the carver formula Fotr risti ("Fot carved"). Here the commissioners and/or the carver of the stone may have considered it sufficient to employ the cross only once. It is of interest that the cross is found together with the carver formula. In the meantime, the nature of the evidence does not allow us to draw any conclusions as to who decided upon its inclusion. In other cases, both monuments in the pair have crosses; various designs can be used. 22 It has been discussed whether crosses made up original elements or were added later (e.g. Kitzler Åhfeldt 2002:44ff). In the case of designs that display the cross separately from the runic carving, it would have been easy to add the ornament after the carving of the inscription was completed. The process of producing rune-stones consisted of several stages; as is known, different people could execute the carving of the runes and the ornament. Hypothetically (some) crosses could have been added at a later stage, possibly on a separate occasion. On the other hand, there is nothing particular in the display of crosses on rune-stones that would suggest that they formed later additions. The main contours of the runic text band/serpent band, as well as the main outline of ornamentation, were probably incised first. Preserved stones that carry an empty/unfinished text band provide certain insight into such practices; the rough outline there may already include the cross. 23 Groove analyses of rune-stone crosses have not provided any proof for the hypothesis of added crosses either (Kitzler Åhfeldt 2002).
Different strategies are used in connection with the display of multiple crosses. The crosses can stand together on the same side as the runic text or alternatively one (or more) can appear together with runes, whereas an extra cross is placed on a different side. 24 Some rune-stones have text on two sides, both equipped with crosses 25 -possibly these originate from different occasions. 26 Anne-Sofie Gräslund has proposed that the extra crosses -when differing significantly from the style of the principal cross -were not contemporaneous with the rest of the runic carving. They may have functioned as benediction crosses in a monumental context, providing evidence of religious rituals that preceded the use of consecration crosses in medieval churches (Gräslund 1991:43;Gräslund 1996:32f). The extra crosses may have provided rune-stones with an extended liturgical meaning. Lager has pointed out that consecration crosses in churches would normally have numbered twelve and appeared inside a circle. Her suggestion is that the extra crosses may have served as a kind of "ceremonial blessing" (Sw. ceremoniell invigning), with possible parallels to the consecration of (outdoor) altars (Lager 2002:223

Crosses on Rune-Stones
The use of multiple crosses on rune-stones may originate from specific practices, but due to the lack of contextual information it is hard to reach a definitive conclusion here. There is considerable variation in the number and types of crosses and in the manner in which they are displayed on the monuments. In comparison, it can be mentioned that medieval altars often have five consecration crosses, one near each corner and one in the centre (Cleve 1956:109). Such crosses are normally small and simplified (Rydbeck 1962:460).
There are some 60 rune-stones with more than one cross in Sweden (including some uncertain cases); 27 most of the material comes from Uppland, with instances from Södermanland, Östergötland, Västergötland and Småland. The majority of monuments employ two crosses, displayed either together on one side or on different sides of the stone. There is no evident reason to regard these crosses as specific additions. Some stones have three or four crosses. The Risbyle stone (U 161) has four crosses, as did possibly the lost rune-stone from Bälinge Church (U 1074). The Högby stone (U 893) has five crosses; the principal cross is in the upper part of the stone, and four cross-marks occur at various spots within the ornamental space. This is the closest we can get to a design that may resemble consecration crosses. Five crosses also occur on the base of the enigmatic Sparlösa stone, Vg 119 (usually dated to the 9th century). They are depicted on the side that carries an additional inscription from the 11th century, and they appear together in a ribbon-like pattern above an image of a mask. An additional small cross-mark occurs inside the sail of the ship on a different side of the stone. The iconography, as well as the dating of Vg 119, is debated. It has been suggested that the cross-ribbon may be a later addition. The results of the analysis carried out by Laila Kitzler Åhfeldt indicate that the carver behind the younger inscription may also have carved the cross-ribbon (Kitzler Åhfeldt 2000:116f). In this case, the row of crosses may have functioned as a kind of blessing of the monument that originated from pre-Christian times. On the other hand, Wideen (1955:148) has argued that the repetition of the cross on the Sparlösa stone has depleted the motif of its symbolic value; according to him, the cross-ribbon is merely decorative.
Returning to traditional rune-stones, we observe that crosses on the same stone may differ in style, but they may also appear identical or similar. Various patterns of layout are used that affect the visual prominence of the crosses. Some crosses function as more or less equal (sym-27 In cases when approximate numbers are provided this results from the uncertain nature of part of the evidence, which would require more commentary than possible within the limits of this article. Crosses on Rune-Stones metric or complementary) entities to each other. 28 Others follow a kind of hierarchical set-up regarding size, position and decorative details. Certain rune-stones have one principal ornamental cross and one or several smaller crosses or cross-marks. 29 It is not always easy to determine which one of the crosses is the principal one. U 161, figure 2, has two crosses that could be considered principal due to their size and their position in the centre and on top of the stone, respectively. The remaining two are of smaller size and form an additional pair on top of the stone. Rune-stones with two crosses do not necessarily indicate the primacy of either one. On the Skåäng stone (Sö 33) one cross stands together with the runic text, whereas a second cross appears alone on the other side of the stone. The separate cross could be considered more significant due to its size and position. In the meantime, the cross that is carved together with the runic inscription may have attracted more attention exactly because it is found on the side that has the message in runes.
In comparison, runic grave monuments also use multiple crosses, usually motivated by the overall composition and decorative design of the monument. Normally there is no reason to assume a special purpose behind multiple crosses other than doubling a decorative motif. For instance, cross ornamentation can be used on both gable slabs. 30 One category of medieval grave monument from Västergötland has a pattern of ornamentation in the middle and runic (and Roman script) inscriptions along the edges. These monuments can be equipped with identical crosses both in the ornamental line and at the start and/or the end of the inscription lines. 31 Although most of these cases are not contemporaneous with the rune-stone tradition, they nevertheless serve as an illustration of decorative ornamental purposes that could motivate the inclusion of multiple crosses.
To conclude, the variable circumstances weaken the hypothesis that associates the use of several crosses on rune-stones with rituals of blessing. The notable variation may have instead resulted from stylistic and decorative considerations and possibly also from the preferences on the part of the carvers and/or the commissioners of the stones. The use of multiple crosses may reflect an optional practice of strengthening the visual message of the stone by doubling essential ornamental motifs. Finally, we must keep in mind that imitation may have played a role as well. Various crosses could naturally have been carved by different people, and perhaps even at different times, but this does not mean that the extra crosses would have brought about a shift in the symbolic meaning of the monument. In further studies it is important to analyse each instance of the use of multiple crosses on its own, as no standard practices can be detected in the material. A more individualized approach would also benefit the analysis of other schemes of layout used in connection with rune-stone crosses. The studied material does reveal certain patterns, but individual considerations may alter customary interpretations of the significance of the cross.

RUNIC CROSSES -CONNECTIONS BETWEEN THE VISUAL AND THE VERBAL
The discussion above pointed out some ways in which the cross can be integrated with the runic carving. This raises the question of the interplay between the visual and the verbal. Is it possible that crosses served the additional function of guiding the viewers' attention to particular parts of the inscriptions? Were the created links intentional? Positive replies to these questions would add an extended gestural and symbolic function to the cross. The second part of this article analyses the connections between cross arms and runic text bands and the placement of runes inside crosses.

Links between cross arms and runic text
The interdependence of ornamentation and runic text has been emphasized in studies that examine the "total composition" of rune-stones, as advocated by Anders Andrén (2000). It has been proposed that the placing of the cross carried its own meaning on rune-stones. Following the points made by Andrén, Julie Lund has looked at the position of crosses in the group of rune-stones that relate the good deed of building a bridge. She finds (Lund 2005:122): […] that the deliberate joining of the image of the cross to the name of the deceased or the word 'bridge' was made to emphasize that the deceased, the bridge, or both should be considered as a Christian person or place.
The approaches that study the interplay between text, images and the materiality of the monument have undoubtedly broadened the study of rune-stones (cf. already Jesch 1998). The method of "visual literacy" (Andrén 2000:10) is a relevant principle to apply in the interpretation of runic monuments. Nevertheless, it has to be questioned as to what degree the emerging links between images and texts were indeed intentional, i.e. planned on the part of the commissioners/producers of rune-stones.
It is tempting to associate the placement of the cross with an extended meaning. From the visual perspective it is relatively easy to find examples where the arms of the cross (most often the side arms, but the upper and lower arms can work as well) connect with parts of the surrounding runic text band. To give one example, on the Skylsta rock inscription (U 86) it could be said that the cross points out the verb rista ("carve") on the left and the name of the deceased, Torbjörn, on the right. But can we claim that actual words get highlighted with the help of the cross and that this was (always) a conscious strategy?
For the analysis of observable links between crosses and elements of runic texts, a corpus was extracted from the Samnordisk runtextdatabas (version 2008); the material was mainly studied on the basis of photographs/drawings available in the database, the corpus edition of Swedish runic inscriptions (Sveriges runinskrifter 1900-) and the journal Fornvännen; this was complemented with a personal investigation of rune-stones. Included in the analysis were rune-stone crosses that are linked to the surrounding runic text by their arms (one arm or several). From the areas that are in focus in this study, Södermanland and Uppland, the analysed cases number 85 and 374, respectively. The material from these two provinces was complemented by consulting relevant inscriptions from other runic regions of mainland Sweden.
This material does not corroborate the idea of intentional connections being created between cross arms and the runic text band. What we find are rather the accidental results of rune-stone layout, depending upon chosen formulations, design and the physical-material features of the monument. The words that are connected to cross arms vary from inscription to inscription and do not follow any evident patterns; in other words, they do not necessarily concern the deceased or the monument. Cross arms can be linked to different nouns, pronouns, verbs, adjectives, prepositions and conjunctions, and they can connect with the names of various people (family members and others) -thus not following any principles regarding a supposed Christian commemoration of the deceased (cf. Lund 2005:122). Secondly -and this is perhaps even more important -the claims that cross arms are attached to particular words are in the majority of cases rather arbitrary. When we start identifying systematically the concrete runes that have an immediate connection with the cross, the links in fact only involve parts of words or just a few random runic signs. In the case of the rune-stone from Frustuna Church (Sö 10), for example, one could claim that the left cross arm points out the name of one of the commissioners (suiunkR, SvaeinungR), and that the right arm connects with the name of the deceased (iarl, Iarl). What we actually see is that the left arm connects with the runes s and possibly u, the right arm with r and l -these spots result naturally from the design of the inscription.
Cross arms can also point out word dividers and empty spaces. They may cut into the runic text and separate parts of the inscription instead of highlighting them. Another aspect to consider is that crosses serve in a similar manner in the so-called non-lexical inscriptions, that is, inscriptions without obvious lexical meaning, earlier labelled "nonsense inscriptions" (on the material of non-lexical inscriptions, see most recently Bianchi 2010:165ff). Monuments with non-lexical inscriptions can be said to imitate the layout of ordinary rune-stones. They mediate the idea of a rune-stone and in doing so they use designs that were considered customary. Marco Bianchi (2010:209) has pointed out that the inclusion of the cross is a common practice on non-lexical rune-stones.
The possible demonstrative and gestural function of the cross is still a relevant concept to consider in the analysis of rune-stones. Although the material does not show that cross arms were deliberately linked to specific parts of the runic text, it remains a possibility that the visual experience of the monument (i.e. reception) would have been influenced by the position, shape and size of the cross as well. In the visual-textual analysis of runic monuments, it is important to consider two dimensions -the first has to do with possible deliberate planning that guided the composition of monuments (i.e. the production phase), the second concerns the circumstances arising when people experienced the monument (i.e. the reception phase).

Runes inside the cross
In addition to links between cross arms and runic texts we should discuss the placement of parts of the runic text inside the cross and the distinctive ways of organizing sequences of runes around the cross. On more than 50 rune-stones in mainland Sweden crosses have some runes carved inside them (cf. group G in Lager 2002:72, 113). Crosses with inserted runes are most common in Uppland and Södermanland, but they also occur in Östergötland, Småland, Västergötland, Västmanland, Gästrikland and Hälsingland. A few parallels are known from Öland, Gotland and Denmark. It should be mentioned that late medieval grave monuments from Gotland also demonstrate the same usage (on the use of crosses on Gotlandic grave monuments, cf. also Staecker 2004). The practice was thus not a limited occurrence, associated with particular places, periods and individuals, although we can observe recurring use Crosses on Rune-Stones Figure 3. Sö 36, Trosa bridge. The word sinn is carved inside the cross. Photo by Kristel Zilmer. in the inscriptions of some carvers. Crosses with inserted runes are, for example, known from the works of Tore in Södermanland and Visäte in Uppland. To this material we can add rune-stones on which the whole inscription or part of it is designed in the form of a cross. 32 Further additions consist of over 30 rune-stones where runes appear inside a staff or a band that is attached to the lower cross arm. 33 But how much awareness existed behind such use of crosses? Do the recorded cases reflect conscious strategies or accidental results of layout? It is logical to expect that the inscription would continue inside the centrally positioned cross when the scheme of layout favoured this and/ or when not all of the text could be fitted inside the text band/serpent. This concerns, for instance, the previously mentioned cross-band stones on which the cross is an integrated part of the inscription. One example is the rune-stone by Trosa bridge (Sö 36) where the possessive pronoun sinn, which concludes the memorial formula, is carved inside the lower cross arm, figure 3. The inscription found its natural conclusion inside the cross; we would be over-interpreting this scheme of layout if we argued that the placing of sinn was deliberate because exactly this word had to be highlighted. On the other hand, even such (accidental) uses cast light upon the flexible attitude towards the role of the cross. The cross did not have to remain an isolated symbol ornamentally or textually; one could use it in different ways. Once the cross was included on the stone, it could be made to work together with the rest of the carving if the occasion so demanded. Further examples of the same functional attitude are inscriptions in which runes are not fitted inside the cross but simply run through it. 34 In order to determine the degree of awareness behind the use of runic text inside the cross, we should explore what it is that is carved there. The clearest evidence of actual consciousness of the presence of the cross is found in two inscriptions where the very word kross ("cross") is carved inside the staff that is attached to the cross. 35 The graphic depiction of the cross has been identified in explicit terms; this seems to be a planned stage.
Of significance are such cases in which the verbal religious message of the runic inscription -i.e. the prayer formula -is carved inside the cross, figure 4. On such occasions the link between the visual and textual is strongly manifested. In the meantime, the components of the prayer 32 E.g. Sö 72 †, Sö 160. Cf. Vg 124. 33 Cf. e.g. Sö 170,Sö 219,U 61,U 316. 34 E.g. Sö 198. 35 Sö 227,Sö 340 (fragmentary).  may be casually divided between the text band/serpent and the cross. The transferring of the prayer to the cross can occur at completely accidental points. 36 Sometimes only the very end of the prayer stands inside the cross; on the Torsätra stone (U 613), we find the word hans ("his") in this position.
Similar non-rigid strategies are visible in other components of runic texts. On the one hand, complete formulations can be fitted inside the cross. On the other hand, the text inside the cross can form a continuation or a short ending of a statement that started in the runic band. Another point to underline is variation; crosses are not reserved for particular types of statements but provide a suitable setting for different textual elements.
One option is to place the complete carver formula inside the cross. 37 The statement inside the cross may be brief, like the carver's name Åsbjörn on the Sanda stone (Sö 266). The manner of highlighting the carver formula by carving it inside the cross seems to be a recurring feature in the inscriptions signed by Visäte; one example is the Lindö stone from Vallentuna (U 236). 38 Besides prayers and carver formulas, crosses contain various additions to the memorial formula, such as additional information about the deceased. The Nälberga stone (Sö 170) has the characterization of the deceased as þrottaR þiagn ("a þegn of strength") inside the elongated staff that is attached to the lower cross arm. 39 Another possibility is to let the cross carry the end of the memorial formula. 40 Occasionally, the inscription may also start inside the cross. Sometimes only a rune or two has made their way into the cross. 41 A different strategy is to arrange runes (partly) around the cross. On some rune-stones the resulting design speaks of some planning (cf. Andrén 2000:20). For instance, on the Lifsinge stone (Sö 9) the prayer Guð hialpi salu Ulfs ("May God help the soul of Ulf") is carefully fitted in between the arms of the cross, figure 5. 42 Other ways in which crosses can be connected with runes include cases when cross arms are used as supporting staves for various types of runic signs; furthermore,  there also occur some cross-like formations made up of runes (cf. Bianchi 2010:121, 125).
Most of these examples show that the carving of runes inside the cross was not a practice of automatic significance, with the possible exclusion of such cases when complete formulations (prayers, carver formulas) have been given this position. Runes could be inserted into the cross when deemed necessary. Besides being a visual symbol with potentially strong external connotations, the cross could thus be employed for text-internal purposes. In connection with this, the cross emerges as a multifunctional entity.

CONCLUSIONS
This article has examined the functions of rune-stone crosses from a contextual perspective that takes into consideration the overall composition of the monument. The visual analysis of the position of the cross has underlined its role as a significant and independent visual marker that communicated a message of its own. The cross would in many cases enjoy a central position on the monument; often it would be visually separated from the rest of the carving or made to stand out in some way. The cross could also appear on a separate side of the runic monument. Such contexts of usage support the idea of the cross as an externalized marker of Christian connections.
In the meantime, there are also numerous cases where the cross is closely incorporated into the general design of the carving and integrated with the text band/runic serpent and/or various elements of ornamentation. Such designs demonstrate the ways in which the cross could be made to serve the internal purposes of the inscription. Decorative and stylistic motives behind the use of crosses have to be taken into consideration.
The visual analysis of crosses has been expanded by the inclusion of the textual dimension, with the purpose of exploring links between crosses and formulations in runes. The demonstrative function of the cross may have played a role in the visual reception of the monument, but it is harder to make concrete claims about intentional planning and conscious layout strategies. The connections that occur between the cross arms and various parts of the runic text seem to be largely accidental and result from the layout of the inscription on the stone; no obvious patterns and underlying principles can be detected. However, certain awareness of the role of the cross can be observed in such cases when specific parts of the inscription (prayers, carver formulas) are placed inside the cross. The ways of organizing parts of the runic text around the cross reflect design-related planning.
It can be concluded that the significance of the cross varied and that there existed different functional attitudes to its role in the composition of runic monuments. Its various functions did not have to exclude each other; they could be made to work together. The cross as the marker of Christianity -and the most visual expression of personalized faithcould simultaneously express more monument-based and inscriptionrelated purposes. It could have an internal decorative and practical motivation that was not directly connected to expressing a message of religious adherence.
Previous research has shown that on earlier rune-stones crosses were more pronounced Christian elements, but their gradual integration with Scandinavian zoomorphic ornamentation downplayed their meaning. Such general tendencies that build upon the stylistic dating of runestones are important to consider, since they demonstrate the changing functions of cross ornamentation. At the same time, we also have to consult different forms of contextual evidence. When we add the textual dimension and explore the use of runes inside crosses, we find that similar strategies could be employed on rune-stones from different parts of Scandinavia; they are even present on some considerably younger medieval grave monuments. Certain functional and practical attitudes towards the role of the cross existed in the monumental tradition from early on, and they did not necessarily disappear together with traditional rune-stones either.
In further studies it would be important to compare the use of crosses on rune-stones in more detail with cross ornamentation on contemporaneous and younger runic and other grave monuments. Recent studies have shown that the rune-stone and the early Christian grave monument traditions related to each other in meaningful ways. Cecilia Ljung has characterized these two types of monuments as variations within the same memorial tradition (Ljung 2009a:595f;Ljung 2009b:195f). This article noted a few parallels in cross-depiction practices on rune-stones and runic grave monuments. Further analysis can bring in additional evidence of the cross as an element of visual expression that could be adapted and modified according to various contexts of usage.