Cover image: Selected examples of Viking Age dog-collar and dog-leash fittings discussed in the article. Photographs: Kirsten Helgeland (UniMus catalogue) and Arnold Mikkelsen (National Museum of Denmark), CC BY-SA 4.0.
Dogs were a visible and socially significant element of Viking Age society, as demonstrated by archaeological evidence from graves, settlements, and iconography. Their presence has often been discussed in relation to social status, hunting practices, and broader human–animal relationships. In contrast, the material culture associated with the physical restraint and control of dogs has received far less systematic attention. Artefacts interpreted as collars, chains, leashes, or related fittings occur sporadically in the archaeological record, and their characteristic features have rarely been considered as a distinct analytical category.
This article focuses on objects that have been interpreted as dog collars or leashes, with particular emphasis on material from high-status burial contexts in Viking Age Scandinavia. By assembling and reviewing this corpus, the study aims not only to describe the formal characteristics of these artefacts, but also to reassess their functional interpretation through close attention to their archaeological contexts and associations with animal remains. Objects traditionally identified as canine equipment are re-evaluated alongside morphologically similar artefacts commonly classified as horse gear, thereby questioning some established interpretations and suggesting alternative functions for these objects.
By comparing material from Norway, Sweden, and Denmark, and by integrating earlier scholarship with new analytical observations, the article seeks to clarify what can and cannot be reliably identified as canine equipment in the Viking Age. More broadly, it addresses the methodological challenges involved in distinguishing between dog equipment and horse gear in Viking Age burial contexts. The study does not aim to establish a definitive typology of dog collars or leashes, but rather to define a set of criteria and limitations that may inform future analyses of this material category.
Introduction
Dogs were a well-established and socially visible element of Late Iron Age and Viking Age society in Scandinavia. Archaeological evidence, particularly from burial contexts, shows that they were regularly included in graves from the Migration Period onward, and this practice continued throughout the Viking Age. Dogs occur in both cremation and inhumation contexts and are documented in a variety of burial settings, particularly high-status graves.
Their role in mortuary practice has been interpreted in different ways. On the one hand, zooarchaeological evidence suggests that dogs fulfilled practical functions in life, including hunting and guarding. On the other hand, their repeated inclusion in graves has been linked to symbolic and ritual dimensions, including associations with protection, liminality, and the transition between life and death. The presence of dogs has also been seen as a marker of social identity, especially in contexts where they accompany other high-value animals.
In elite ship burials such as Gokstad, where horses, hunting birds, and dogs occur together, the animal assemblage has been interpreted as reflecting aristocratic hunting practices and the social position of the deceased. The combination of these species suggests access to trained animals and participation in specialised forms of elite hunting rather than simply indicating the presence of domestic companions (Gansum 2018).
Material from the Gokstad and Oseberg Ship Burials
Against this broader background, the ship burials of Gokstad and Oseberg occupy a central position, dated respectively to c. 900–905 and 834 (Bonde and Christensen 1993). Both graves combine the intentional deposition of dogs with exceptionally rich assemblages of animals, equipment, and fittings, including extensive horse gear. At the same time, they provide one of the largest and most coherent sets of artefacts that have been interpreted as elements of dog collars, leashes, and related restraint devices. The coexistence of dogs and horses within the same burial contexts, together with the formal similarity between fittings associated with these animals, makes Gokstad and Oseberg particularly well suited for examining questions of functional attribution. These sites therefore form the starting point for a closer analysis of artefacts traditionally identified as dog equipment, with special attention paid to form, construction, and archaeological context.
Gokstad
In the Gokstad mound, Nicolaysen (1882) recorded the remains of six dogs; however, more recent osteological analyses indicate that there were at least seven (Gansum 2018). Beside two dogs lying outside the stern of the ship, fittings were discovered that have been interpreted as belonging to collars (Nicolaysen 1882, 52).
One of these (C10430) consists of two folded bronze plates, approximately 2.1 cm in width and from 3.2 to 3.6 cm in length, resembling plates known from belt buckles, each with a bronze rod set through the centre. These are joined by an oval-shaped iron ring measuring 7.2 cm in length and 2.2 cm at its widest point (Fig. 1A). One of the strap mounts bears a short runic inscription (Fig. 1B), incised shallowly and with limited care. The reading and meaning of this inscription have not yet been examined in detail.
Fig. 1. Dog-collar fitting from Gokstad (C10430). (A) General view. (B) Enlarged detail of the runic inscription on one of the strap mounts. Photograph: Kirsten Helgeland, UniMus catalogue, CC BY-SA 4.0.
Near the second dog skeleton lying outside the stern of the ship, a similar object was discovered; in this case, however, it was made entirely of iron (C10431) (Fig. 9A) (Nicolaysen 1882, 52). In this example, the oval ring narrows slightly in its central section. The strap mounts are of simpler construction than in the preceding specimen: they are forged from iron and each is fastened with two iron rivets. They are also markedly narrower, measuring approximately 1 cm in width, and considerably longer, at around 6 cm in length, than the corresponding element of the collar with bronze strap mounts.
Together with this object, a fragment of chain measuring 16 cm in length was found, composed of figure-of-S-shaped links, of which three complete and one incomplete example are preserved (C10429) (Nicolaysen 1882, 52).
Inside the sepulchral chamber, an object closely resembling the presumed collar element made entirely of iron (C10431) was also found; in this case, however, it included an additional circular ring attached to the oval ring (C10435) (Nicolaysen 1882, 48).
In the Gokstad mound, a bronze swivel (C10822) (Fig. 5A) was also recovered; however, this find derives from an unspecified lump of clay from the mound, and it is therefore not possible to determine the precise location within the burial from which it originated.
Oseberg
The Oseberg mound contained more grave equipment than the Gokstad mound. The burial included the remains of four dog skeletons and fifteen horse skeletons (Christensen, Ingstad and Myhre 1992).
Among the assemblage of metal fittings recovered from the burial was an object closely comparable to the find from Gokstad that has been interpreted as a fragment of a dog collar (C10430). The Oseberg example (Fig. 2A) likewise consists of two bronze strap-mount plates and an oval iron ring connecting them, together with a circular ring attached to the oval ring (C55000/UN3). The dimensions of the oval ring, measuring 6.5 cm in length and 2.3 cm in width, are comparable to those of the Gokstad example. The circular ring measures 2.2 cm in diameter. In this case, the strap mounts are wider, approximately 3.1–3.2 cm in width, roughly square in shape, and neither bears a runic inscription.
Fig. 2. (A) Dog-collar fitting from Oseberg (C55000/UN3). (B) Twenty tin fittings, an iron belt buckle, and an iron swivel interpreted as belonging to a dog collar from Oseberg, mounted on a modern reconstruction. Photographs: Kirsten Helgeland (A and B), UniMus catalogue, CC BY-SA 4.0.
In addition, a group of twenty small decorative mounts made of a tin alloy was recorded from the Oseberg burial. Each has a convex boss surrounded by a beaded rim and measures between 0.8 and 0.9 cm in diameter (C55000/200.2). These mounts are currently mounted on a reconstructed example of a presumed dog collar prepared for museum display, and their original arrangement cannot be established with certainty. The present reconstruction also incorporates a small iron buckle and an iron swivel (C55000/83) (Fig. 2B).
Four further objects were recovered at Oseberg (C55000/102; C55000/243.d; C55000/69; C55000/68), each consisting of an oval ring and two iron strap mounts, three of them also incorporating a circular ring passing through the oval ring. These closely resemble two objects from the Gokstad mound, one of which (C10431) was found beside a dog skeleton and is interpreted as a dog collar fitting, while the other (C10435) was recovered in a context suggesting use as horse gear.
Two of the four objects, of very simple construction and made entirely of iron (C55000/68 and C55000/69) (Fig. 9B), were found on horse skeletons on the starboard side of the Oseberg ship. Their find context therefore supports an interpretation as elements of horse gear. The remaining two objects (C55000/102 and C55000/243.d) (Fig. 9C) are very similar to one another and differ from the previously discussed pair in that their iron strap mounts, measuring from 5.1 to 6.0 cm in length, are fastened with two tin rivets. Moreover, these iron strap mounts are decorated with between one and four simple incised lines located between the first tin rivet and the loop-forming section bent around the oval ring. The oval rings of both objects measure from 6.6 to 6.7 cm in length and from 2.1 to 2.6 cm in width, which closely correspond to the dimensions of the rings belonging to the presumed dog collar fittings from Gokstad, as well as to the Oseberg example with bronze strap mounts (C55000/UN3). The circular rings measure 4.0 and 2.9 cm in diameter respectively. Although two of these objects, interpreted on the basis of their find context as horse gear fittings, possess oval rings of comparable dimensions, these rings are essentially the same size as those forming part of fittings interpreted as elements of dog collars.
The Oseberg mound also contained several iron chains relevant to the present analysis.
One of these chains (C55000/129.I) can be interpreted as a cauldron suspension, while several others were found in association with dog skeletons and have therefore been interpreted as dog leashes (Grieg 1928). One such example (C55000/200.1) (see Grieg 1928, fig. 158b; Fig. 3) currently measures 108.7 cm in length. It is divided into three sections and consists of thirty-three links, together with a circular ring at one end and a clearly identifiable fragment of an iron swivel preserved at the opposite end. Thirty-one of the links are of simple construction, each made from an oval ring folded at the midpoint. By contrast, the terminal and initial links are made from oval rings clamped at the midpoint with a cube-like element, giving the link a figure-of-eight appearance. The surface of these terminal links is decorated with a lattice pattern composed of small engraved lines. The chain is generally interpreted as having originally functioned as a dog leash (Gustafson 1904a, 534). Given its present fragmentation, resulting from missing links or longer sections, it may be assumed that the chain was originally longer than its currently preserved length.
Fig. 3. Iron chain from Oseberg (C55000/200.1), interpreted as a dog leash, composed of simple links. Photograph: Kirsten Helgeland, UniMus catalogue, CC BY-SA 4.0.
Another chain (C55000/66.a) (see Grieg 1928, fig. 158a) was found in the fore part of the ship, next to a dog skeleton, behind Shetelig’s sledge. It currently measures 96.5 cm in length and consists of twenty-two links. In addition, two smaller detached sections of the chain are preserved, consisting respectively of two and one link, as well as five small link fragments. Its original length must therefore have been greater than that presently preserved. All of these links are of the same simple construction as the thirty-one links of the chain described above.
A further chain fragment differs from the others (C55000/67) (see Grieg 1928, fig. 158c). It was found beneath a horse skeleton in the fore part of the ship. It presently consists of five links, one of which, presumably the terminal link of the chain, is broken and survives in two parts. These links resemble the two distinctive terminal links of the first Oseberg chain described above, which formed its beginning and end. The figure-of-eight-shaped links are made from oval rings clamped at the midpoint with a polygonal iron block; in this case, however, they lack any form of ornamentation. The surfaces of the links display areas of bright, silvery and golden appearance above a reddish-brown corrosion layer (Fig. 4). XRF analysis identified iron, copper and zinc, suggesting that the iron chain was originally coated with brass (Frantzen 2021, UniMus online catalogue, accessed 12 January 2026). In addition to the five links, the chain comprises an oval ring and a circular ring; the entire preserved assemblage currently measures 34.2 cm in length. The oval ring measures approximately 5.2 cm in length and 2.0 cm in width, while the circular ring measures approximately 2.5 cm in diameter. These rings closely resemble the paired ring arrangements observed on objects interpreted as elements of dog collars and, in other cases, as fittings presumably belonging to horse gear.
Fig. 4. Chain fragment from Oseberg (C55000/67) with links clamped by a polygonal iron block. Photograph: Kirsten Helgeland, UniMus catalogue, CC BY-SA 4.0.
In addition, several smaller chain fragments were found at Oseberg. These include a fragment measuring 45 cm in length (C55000/101), consisting of only fifteen simple links, as well as another chain fragment now preserved in four separate parts (C55000/195.G.h). The latter consists of eleven simple links, one fragment of a simple link, and one figure-of-eight-shaped link with a clamping polygonal iron block lacking any decoration. The surface of some elements of this chain displays areas of a bronze-like layer, comparable to that observed on the previously described chain (C55000/67). A circular iron ring is attached to the figure-of-eight-shaped link, to which a single iron strap mount is currently affixed. These fragments were found in association with one of the sledges, together with an iron hook with a shaft belonging to a rattle, indicating that this chain was more likely associated with the sledge rather than with animal-tethering equipment.
At Oseberg, an iron swivel (C55000/103.1) was also recovered together with a single figure-of-eight-shaped chain link that probably formed the terminal link of one of the chains.
Analysis
The archaeological material from the Gokstad and Oseberg mounds provides a substantial body of evidence for identifying different types and forms of fragments that may be interpreted as belonging to dog collars and dog leashes. This material is crucial for isolating and defining their characteristic features. The discussion begins with dog collars.
In the case of the Gokstad collar fitting bearing a runic inscription, this interpretation appears well supported by its context and thus provides valuable typological information concerning Viking Age dog collars. The object consists of two strap mounts made from folded bronze plate, which in this case were most likely attached to a leather collar by means of between three and six rivets. The two bronze strap mounts are connected by an oval ring. It appears that this collar fitting lacks the smaller circular ring present on the closely comparable object from Oseberg, whose strap mounts are likewise made of bronze plate and fastened with four rivets each (Fig. 2A).
The function of such elements can now be considered. The most plausible interpretation is that they served a tightening function. When a dog leash is attached to the circular ring and the dog pulls forward, the circular ring slides along the oval ring, which rotates by approximately 90 degrees. This movement causes the entire collar to tighten, drawing the two strap mounts closer together. Such a mechanism serves several purposes. First, the substantial reduction of the collar’s circumference decreases the risk of the dog slipping its head out of the collar during sudden or uncontrolled movements. Second, the mechanism has a behavioural effect: as the collar tightens around the carotid area and exerts pressure on the trachea, it causes pain and partial choking, prompting the dog to yield and become easier to control. This function is particularly relevant in the case of dogs trained using aversive methods and dogs employed in hunting.
The dogs buried with the Gokstad chieftain were likely employed for such purposes, as hunting formed an integral part of his activities during life (Gansum 2018). Birds were also employed in hunting, as evidenced by the presence of two female goshawks in the mound. The breeds of the buried dogs have been identified as well, comprising two elkhounds and five greyhounds (Gansum 2018). A dog of the latter breed is likely depicted on the lid of the Gokstad backpack, an object generally regarded as hunting equipment (Hinsch 1958). The carvings on this lid probably depict a hunting scene in motion, a motif that is comparatively rare in Viking Age art (Blindheim 1981). The animal shown pursuing forest prey, possibly a deer, closely resembles a greyhound, particularly in the distinctively slender head, proportionally small in relation to the body and strongly extended forward during the chase.
Collars with a tightening function, such as those from Gokstad and Oseberg, were already in use prior to the Viking Age, during the Merovingian and Vendel periods. A collar with a tightening mechanism made of iron was found together with the remains of two dogs in the boat grave Valsgärde 10 in Sweden (Öhman 1983, fig. 8). This collar was additionally equipped with four iron mounts, each fitted with four spikes interpreted as protecting the dog’s neck. Such collars continued to be used in Scandinavia until the late nineteenth century to protect hunting, herding, and guarding dogs (after Öhman 1983, 174). Another example derives from the Vendel III grave (Öhman 1983, fig. 7a), where two folded iron plates similar to those of the tightening collar from Valsgärde 10 were found in association with dog remains. A comparable tightening mechanism, probably belonging to a dog collar, is also known from Norway; however, its findspot and archaeological context are unknown. The object is preserved in the collections of the Kulturhistorisk Museum in Oslo under the inventory number C22988.
A second type of dog collar may be represented by collars lacking a tightening function but characterised by a more decorative appearance, such as the presumed collar from Oseberg fitted with twenty circular tin mounts (C55000/200.2), a small iron buckle, and an iron swivel (Fig. 2B). It is not certain, however, that all of these elements originally belonged to the same object, nor that the object in question was indeed a dog collar. Even more uncertain is whether the iron swivel formed part of the same item as the tin mounts and the iron buckle, as these components are listed under entirely different entries in Gustafson’s excavation diary (Gustafson 1904a, 55; 534–535). If it is assumed that the twenty circular tin mounts together with the iron buckle originally constituted a dog collar, such a collar would most likely have served a primarily representational function. This interpretation follows from the absence of a tightening mechanism and from the enhanced visual prominence provided by its decoration. In the case of tightening collars, no buckles corresponding in width to their strap mounts have been identified, comparable to the iron buckle attributed to the Oseberg example. It must therefore be assumed that tightening collars were not equipped with buckles of the type known from waist belts or horse harness. Such buckles were unnecessary, as collars that tightened under tension were already difficult to remove and did not require adjustment of their circumference. The situation differs in the case of collars lacking a tightening mechanism, such as the presumed Oseberg collar with tin mounts. In such cases, the ability to adjust the circumference would have been essential, which provides a plausible explanation for the presence of an iron buckle.
Swivels
In the modern reconstruction of the presumed Oseberg collar prepared for museum display, a small iron swivel, that is, a rotating connector, has also been mounted. A swivel fragment of slightly different form is also preserved at one end of a chain from Oseberg that has been interpreted as a dog leash (Fig. 3). In addition to these examples, a further swivel fragment of similar form was discovered in the Oseberg mound (C55000/103.1), together with a single link of the same type as those found on one of the chains interpreted as a dog leash. This component plays an important functional role within a collar and leash set. By allowing rotation through 360 degrees, it prevents the leash from twisting and tangling, compensating for the torsion generated by the movement of the animal and the leash. A swivel was also recovered from the Gokstad mound (C10822) (Fig. 5A). Measuring 3 cm in length, this example was cast in bronze and therefore differs in form from the forged iron swivels from Oseberg. The bronze swivel from Gokstad is not unique. Another cast bronze swivel in cylindrical form (C4312/b) (Fig. 5B) is known from one of the wealthiest male graves at Kaupang in Vestfold, dated on the basis of its assemblage to the ninth century (Blindheim et al. 1981). This swivel measures 5.8 cm in length and 2 cm in diameter and was manufactured from multiple components with a high degree of precision. It most likely served as part of a leash made from a leather strap, as indicated by the flattened terminal to which it could have been attached.
Fig. 5. Bronze swivels from (A) Gokstad (C10822) and (B) Kaupang (C4312/b), Vestfold, Norway. Photographs: Olav Heggø (A and B), UniMus catalogue, CC BY-SA 4.0.
Swivels occur in large numbers in the archaeological material from the Viking Age in Norway. They are predominantly forged from iron and take relatively simple forms (Fig. 6). In the vast majority of cases, they are found either alone or together with one or two chain fragments, and only very rarely in association with complete chains or more substantial chain sections. The most common type is represented by the form illustrated in Fig. 6A, E, and I, in which one of the two components of the swivel resembles a figure-of-eight chain link. Jan Petersen (1951) observed that these swivels occur most frequently in the regions of Østlandet and Trøndelag, a pattern that is confirmed by the analysis conducted for the present study. Petersen’s analysis further indicates that swivels were overwhelmingly associated with equestrian equipment, while largely excluding their functional use in chains serving as cauldron suspensions. This observation is of particular importance for the interpretation of the iron chains discussed below, as it is supported by the frequent occurrence of swivels in male graves containing elements of horse gear, including bits, harness components, and other equestrian equipment.
On the other hand, swivels also formed part of various types of chains and other tethering devices used to restrain animals, such as horses or dogs, as noted above. In this context, a chain from a male grave at Torshov in Akershus (C20168/ø) is of particular interest. The grave contained equestrian equipment and, on the basis of its assemblage, can be dated to the ninth century. The chain measures 1.20 m in length and consists of forty-one iron links of a simple type. It terminates in a swivel connecting it to a broad iron buckle measuring 4.1 cm in width and fitted with a massive iron strap mount. Originally, two buckles were attached to the strap mount, one on each side, although one is now missing. A comparable iron strap mount incorporating a swivel fragment is known from Risan in Trøndelag (T3618), suggesting that tethering devices of this type were more widely distributed. Both finds belong to a group of animal-tethering fittings that were mounted directly onto a leather strap. Comparable examples are known from Holbø, Oppland (C8641); Hemrum, Innlandet (C19699); Dale, Telemark (C25770/z); Egge, Trøndelag (T5074); Ophus Nordre, Innlandet (C1988); Somdalen, Buskerud (C2262); and Myklebost, Sogn og Fjordane (B5807/II_nn). In most cases, however, the substantial presence of equestrian equipment in the graves from which these objects derive makes their interpretation as dog equipment less likely. The objects from Torshov (C20168/ø), Risan (T3618), and Myklebost (B5807/II_nn) merit consideration as a distinct group, as all three are characterised by strap mounts that originally carried two buckles, one on each side.
Fig. 6. Examples of different forms of iron swivels from Norway. (A) Nordgården, Telemark (C17434); (B) Grette, Innlandet (C22061/j); (C) By, Innlandet (C9542); (D) Nordgaarden Nedre, Trøndelag (C583/a); (E) Espen, Innlandet (C21237/r); (F) Ophus Nordre, Innlandet (C1987); (G) Grette, Innlandet (C22785/l); (H) Gjone Søndre, Vestfold (C32023/k); (I) Kopsland, Telemark (C10316). Photographs: Olav Heggø (A, C, F, H, I), Eva Brænd (B, E, G), and Ulla Schildt (D), UniMus catalogue, CC BY-SA 4.0.
Leather-Strap Dog Leashes
As noted earlier, alongside iron chains, dog leashes used in Viking Age Scandinavia were also made from leather straps, as indicated, for example, by the construction of the swivel from Kaupang. Further evidence for the use of leather leashes is provided by dog-leash junction pieces from Denmark. One such object comes from the well-known mound at Ladby in Kerteminde on Funen, which contained an exceptionally rich boat burial (Sørensen 2001; Müller-Wille 1978a; Thorvildsen 1957). The junction piece is executed in the Jelling style and is made of gilded bronze, measuring approximately 7.7 cm in length (see Sørensen 2001, fig. 4.45; Müller-Wille 1978a, fig. 4:9). It consists of two elements that probably originally rotated independently around their own axes, together forming a swivel mechanism. One end of the object is decorated with the characteristic Jelling-style animal heads, arranged symmetrically on either side, between which a leather strap was fastened by means of a rivet. The opposite end likewise bears a pair of animal heads; in this case, however, they are clearly splayed and set further apart, allowing two separate leather straps to be attached with rivets, one to each head. The object was found in the fore part of the ship among the remains of eleven horses and at least three or four dogs (Sørensen 2001; Müller-Wille 1978a), together with other elements interpreted as belonging to dog equipment. These included several bronze oval rings fitted with swivels and circular rings, also made of bronze, as well as two rhomboid decorative plates (Sørensen 2001, fig. 4.45).
In these examples, the oval rings do not have smaller circular rings passing directly through them to which a swivel connecting them to the leash would have been attached. Instead, the oval rings are fitted with very small loops that move freely around the entire ring and connect to a larger circular ring, forming a pivot joint functioning as a swivel. An oval iron ring with a swivel of the same type is known from By, Innlandet, Norway (C9542), as well as from Alm, Innlandet, Norway (C3873/2). These rings likely belonged to a tightening mechanism of a dog collar. In the examples from Ladby, however, the strap mounts known from other examples of this object type are missing, making it impossible to confirm this interpretation conclusively. Alternatively, the oval ring may have moved freely along the dog collar passing through it, although such a solution would appear to be of limited practical functionality. One possible reconstruction of the dog leash from Ladby is illustrated in Müller-Wille 1978a, fig. 10. This type of leash construction would have allowed the handling of four dogs simultaneously. A comparable arrangement is depicted in one of the hunting scenes on the Bayeux Tapestry, shown in the narrow frieze below the main narrative, where a hunter blowing a horn is shown holding a single leash to which two dogs wearing collars are attached.
Fig. 7. Dog leash junction piece from Lille Lihme, Nørup, Denmark. Photograph: Arnold Mikkelsen (National Museum of Denmark), CC-BY-SA 4.0.
A second decorative leash junction piece, fulfilling a function similar to that of the Ladby example, was found accidentally in a field at Lille Lihme, approximately 10 km southwest of Jelling (Thorvildsen 1957). It measures 13 cm in length and is cast in bronze with inlays of silver or niello (Lindblom and Adamsen 2014). The object is currently preserved in five elements, three of which constitute its main body. These components are arranged in such a way that they are able to rotate independently through 360 degrees around their own axes, together forming a swivel mechanism. The central part consists of two bird-shaped figures, each depicted with the head turned backwards. On either side of the central element, a loop is attached; one of these loops is decorated with two animal heads in the Jelling style. To one of the loops, a strap mount is attached, which originally connected the object to a leather strap by means of a bronze rivet that is still preserved in place. The strap mount is U-shaped and terminates at both ends in animal heads in the Jelling style. An identical strap mount is attached to the other loop of this leash junction piece; however, in this case one mount is apparently missing, as indicated by wear visible at two points on the loop. In this instance, the object represents a further junction piece branching into at least two leashes, and another elaborately decorated element of a leather-strap dog leash from Denmark, demonstrating a high level of craftsmanship.
Iron-Chain Dog Leashes
As discussed above with reference to the material from the Oseberg mound, several variants of iron chains can be interpreted as dog leashes. These include iron chains composed of very simple links made from folded oval rings, as well as examples with figure-of-eight-shaped links clamped either with an ornamented cube-like element or with a polygonal iron block. Chains of this type may have been coated with brass, as demonstrated by the analysis of the Oseberg chain registered under inventory no. C55000/67. Iron chain leashes were fitted with a swivel at one end, usually made of iron. It cannot be excluded that some bronze swivels lacking a clear attachment point for a leather strap, such as the example from Gokstad, were instead intended for use with iron chain leashes.
Iron chains that can be interpreted as dog leashes also occur in the Swedish archaeological material. The male boat burial Valsgärde 14, dated to the early ninth century, represents one of the earliest Viking Age weapon graves in the Mälar Valley (Nordahl 2018). Two iron chain dog leashes were recovered at this site in association with two dogs intentionally buried with the deceased (see Nordahl 2018, figs. 22–23). In this case, however, the dogs were preserved only as faint soil discolorations. Both leashes measure slightly over one metre in length and consist of twelve oval links connected along almost their entire length. At one end of each chain there is a swivel, while at the other end a folded iron mount served as the attachment point for a leather thong leading the leash to the hand (Nordahl 2018). A comparable dog leash was also found in the boat burial Valsgärde 28, dated to the Vendel Period (Nordahl 2018). This example consists of seven simple links, with an iron swivel terminating in a ring at one end and a folded iron mount at the other (see Nordahl 2018, fig. 24).
Further examples of iron chain dog leashes are known from Sweden, including the boat burial at Nabberör on Öland. On the basis of artefact typology, the Nabberör grave has been dated to the Late Vendel or Early Viking Age (700–800). The burial contained the remains of between ten and twelve individuals and five large dogs, three of which were interred with iron leashes (Magnell, Iregren and Jennbert 2023, fig. 15). Two of these leashes, from clusters IV and IX, are composed of numerous links of a simple type. The leash from cluster IV preserves one element of a swivel at one end, while the example from cluster IX is likewise fitted with a swivel at one end and a folded iron mount attached to a ring at the other. The third leash from Nabberör, found in cluster II, consists of at least five links made from elongated iron rods. One end terminates in an iron ring fitted with a folded iron strap mount, whereas the opposite end is equipped with a swivel. A small iron buckle was also found in association with this leash and may have belonged to a leather dog collar.
In the context of Swedish iron chains that may have functioned as dog leashes, the material from the cemetery at Vendel also merits consideration. Iron chains are recorded from the boat graves I (Öhman 1983, 170–171), III (Öhman 1983, fig. 7b), VI (SHM, Stockholm, inv. no. 3121262), VII (Öhman 1983, 171), VIII (SHM, Stockholm, inv. nos. 3236255 and 3236256), and XII (Öhman 1983, 170). These iron chains, found in connection with dog remains, are likewise made of iron and are fitted with iron swivels at one end; in some cases, the opposite end carries a circular iron ring and a folded iron strap mount. They are generally composed of simple links formed from folded oval rings. In certain examples, such as the chain interpreted as a dog leash from grave VI, one link is clamped with a polygonal iron block. Fragments of an iron chain with a circular ring and a possible folded iron mount are also known from grave 944 at Birka (Arbman 1943, 368–371; see Tafeln 30). No dog remains, however, were recorded in that grave. Another iron chain that may be interpreted as a possible dog leash derives from the Mästermyr find on Gotland (Arwidsson and Berg 1983, 8–9, 27; pl. 16:17).
These examples, dating largely to the Vendel Period and the early Viking Age, provide valuable information regarding dog leash construction in this period. The tradition appears to have continued at least into the late Viking Age, as indicated by later examples from Norway. In this context, the chain from the elite male burial at Gjermundbu in Ringerike deserves closer examination. The grave is dated to the second half of the tenth century, most likely to its final decades (c. 970–990), on the basis of weapon typology, in particular a Petersen type S sword decorated in the Jelling style, as well as late Viking Age spearheads (Petersen 1919; Stylegar and Børsheim 2021). A chain was recovered from the burial that Grieg (1947) interpreted as a cauldron hook. As first suggested by Else Nordahl (2018), however, the construction of the object suggests that it may have functioned as a dog leash. The chain is composed of simple links of a type well known from other dog leashes. One end consists of a simple swivel fitted with a large iron ring. At least ten links further along, the chain, which now survives in several fragments, incorporates a simple divider made from an iron rod that branches the chain into two sections. A comparable triple iron divider, interpreted as part of equestrian equipment, was found in one of the boat burials at Borre, Vestfold, which also contained the remains of a dog and horses and has been dated to c. 850–900 (Myhre 2015, fig. 2). Each branch consists of seven simple links terminating in swivels similar to that at the beginning of the chain. This configuration recalls the branching dog leashes from Ladby and Lille Lihme, which likewise divided into at least two leads. In the case of Gjermundbu, however, despite the equally high-status character of the burial (Stylegar and Børsheim 2021), the construction of the leash is relatively simple.
In this context, it is appropriate to consider whether other chains from Norway might warrant reinterpretation in light of the structural characteristics typical of iron chain dog leashes. One such example is the chain from the Oseberg mound registered under inventory number C55000/195.G.h. This chain terminates at one end in a folded iron mount attached to a circular ring, closely comparable to the construction observed in the dog leashes from Nabberör and Valsgärde (Magnell, Iregren and Jennbert 2023; Nordahl 2018). It is also composed of simple links of the type characteristic of iron chain dog leashes. Its function as a dog leash may therefore be considered, although the burial context in which it was found does not directly support such an interpretation and leaves open the possibility that it served to restrain another animal. Further Norwegian examples that merit consideration in this regard include chains from Fevang Nordre, Vestfold (C6483); Dale, Telemark (C25770/z); Bygland, Telemark (C27454/xxx.I); Kåsen, Telemark (C21791/m); Holbø, Innlandet (C8639); Heen Nedre, Innlandet (C29600); Moksnes, Trøndelag (T2053); Nordgaarden Nedre, Trøndelag (C584/a); and Hennum, Buskerud (C5485) (Fig. 8). These examples share features typical of iron chain dog leashes, including simple links formed from folded oval rings, the presence of at least one swivel, and in some cases a folded iron strap mount at one end. Such characteristics, however, cannot be regarded as conclusive evidence of their original function, particularly as many of the graves from which they derive also contained elements of equestrian equipment.
Fig. 8. Examples of iron chains from Norway displaying morphological features that may indicate their use as dog leashes. (A) Fevang Nordre, Vestfold (C6483); (B) Moksnes, Trøndelag (T2053); (C) Bygland, Telemark (C27454/xxx.I); (D) Dale, Telemark (C25770/z); (E) Kåsen, Telemark (C21791/m); (F) Holbø, Innlandet (C8639); (G) Heen Nedre, Innlandet (C29600); (H) Hennum, Buskerud (C5485); (I) Nordgaarden Nedre, Trøndelag (C584/a). Photographs: Brynjar Sandvoll (A), Ole Bjørn Pedersen (B), Olav Heggø (C, D, E, F, G, H), and Ulla Schildt (I), UniMus catalogue, CC BY-SA 4.0.
The previously mentioned chain fragment consisting of three complete and one fragmentary figure-of-S-shaped links may likewise be interpreted as part of a dog leash, as it was found together with an iron tightening mechanism from a dog collar near two dog skeletons at Gokstad (Nicolaysen 1882, 52). Comparable figure-of-S-shaped links are known from grave B4 at Stangå kyrkogård, Gotland (SHM, Stockholm, inv. no. 853362_HST). Chains composed of comparable figure-of-S-shaped links are also known from Somdalen, Buskerud (C2262), and Moksnes, Trøndelag (T2053) (Fig. 8B). If the Gokstad chain originally functioned as a dog leash, the bronze swivel from the same site (C10822) (Fig. 5A) may represent a plausible associated element, as it is fitted with two circular loops, one on each side, to which an iron chain could be attached. Attention should also be drawn to a fragment of an iron chain from Sem, Vestfold (C854), terminating in an iron swivel and presumably dating to the Viking Age. As this is a stray find, its function cannot be determined; nevertheless, the form of its links is distinct from what is typically observed in Norway.
Oval Rings and Tightening Mechanisms
As discussed above, tightening mechanisms of dog collars were produced both from iron rings combined with bronze strap mounts, as in the examples from Gokstad and Oseberg, and entirely from iron, as in the case of Valsgärde 10 and one of the fittings from Gokstad (Fig. 9A). In some instances, the mechanisms also incorporated rings made of bronze, as exemplified by the find from Ladby (Sørensen 2001).
Fig. 9. Examples of different iron fittings interpreted as tightening mechanisms from Norway. (A) Gokstad, Vestfold (C10431); (B) Oseberg, Vestfold (C55000/69); (C) Oseberg, Vestfold (C55000/102); (D) Thingstad Søndre, Innlandet (C4103); (E) Kvarberg, Innlandet (C7667); (F) Tolstad, Innlandet (C18207); (G) Ophus Nordre, Innlandet (C1990); (H) Hemrum, Innlandet (C19694); (I) Leine, Innlandet (C18129). Photographs: Pia Edqvist (A), Kirsten Helgeland (B, C), Olav Heggø (D, G), Maria Malherbes Jensen (E), and Eva Brænd (F, H, I), UniMus catalogue, CC BY-SA 4.0.
An important issue that warrants further consideration concerns the interpretation of elements associated with tightening mechanisms of dog collars, or rather objects whose form frequently resembles such mechanisms. Numerous examples of this type are known from the Norwegian archaeological record. They often consist solely of a pair of rings: one oval and one, usually smaller, circular ring. In many cases, one or two iron strap mounts are attached to the oval ring, typically made from a folded iron plate terminating in a loop through which the oval ring passes. Objects of this kind, occurring in various configurations and with differing numbers of preserved rings and strap mounts, have traditionally been classified as type R.577 and dated mainly to the early and middle Viking Age (Rygh 1885; Petersen 1951). In the course of the present study, a broad range of such objects from Norway has been examined. Among the first examples discussed in this article are two fittings from Gokstad made entirely of iron. One of these (C10431) (Fig. 9A), found near the remains of dogs together with fragments of a chain interpreted as a dog leash (Nicolaysen 1882, 52), may plausibly be understood as a tightening mechanism from a dog collar. The other, a very similar object incorporating a circular ring attached to the oval ring, was recovered from the sepulchral chamber (C10435) (Nicolaysen 1882, 48). Given that the objects found in association with dog skeletons were located outside the sepulchral chamber, whereas a substantial quantity of horse gear was recovered within it, this latter object is more plausibly interpreted as part of the equestrian equipment.
Such a tightening mechanism could therefore have served as a tether for a horse as well as for a dog. At the same time, objects of this type may also be interpreted as a form of rattle within horse harness, particularly since some iron examples incorporate more than one circular ring attached to the oval ring. Examples of this variant are known from Lalum Søndre, Innlandet (C23067/g); Hjemstølen Kvåle Nedre, Innlandet (C23947/q); Finstad, Innlandet (C26405/k); Enelaug, Innlandet (C10726); Østby, Buskerud (C25576/ll and C25576/mm); Modum Prestegård, Buskerud (C3033/d); and Farmen, Vestfold (C13580). Petersen (1951) explicitly states that objects of this kind functioned as connectors or fastening elements within horse harness and does not suggest that any of them served other purposes. On the basis of the material reviewed here, there is little reason to question Petersen’s interpretation. Nevertheless, given the formal similarity between these objects and tightening mechanisms of dog collars, it remains possible that a certain number originally fulfilled such a function. This suggestion, however, should be limited to examples comprising only a single oval ring and a single circular ring.
Summary
The analysis presented here suggests that dog equipment in the late Vendel Period and the Viking Age was neither commonplace nor incidental, but instead appears to have been associated primarily with a limited group of elite burials of a clearly demonstrative character. The archaeological evidence indicates that both dogs and elements of their equipment occur primarily in high-status graves, dated mainly to the ninth and tenth centuries, such as Gokstad, Oseberg, Ladby, and the elite burial from Kaupang. Importantly, the material from Oseberg shows that dog equipment was not confined to male burials, but was also present in female graves of exceptionally high status.
At the same time, in most cases, the burial assemblages examined here combine the presence of dogs and their potential equipment with a broad range of equestrian equipment and, in some cases, horse remains. This coexistence constitutes a central interpretative problem. Objects such as swivels, iron chains, oval rings with smaller circular rings, and fittings with iron strap mounts, in some instances interpreted as tightening mechanisms of dog collars, are formally ambiguous and may have functioned either as elements of dog collars or as components of horse harness. Their mere presence in a grave containing dog remains cannot be regarded as sufficient evidence for identifying them unequivocally as dog equipment, particularly when the same context also includes numerous elements of equestrian equipment or the remains of horses, unless the objects occur in direct association with the dog remains.
As noted by Müller-Wille (1978a), the presence of animals in graves was not a widespread phenomenon but was largely confined to so-called Prunkgräber (elite graves), in which animals formed an integral part of the funerary assemblage alongside weapons, harness equipment, luxury objects, and tableware. In such contexts, dogs were deposited intentionally rather than incidentally, which enhances the significance of their presence but does not in itself resolve the interpretative difficulties posed by the associated artefacts.
The results presented by Else Nordahl (2018) indicate that even in regions where dogs occur relatively frequently in burials, leashes and other restraining devices are only sporadically attested. The absence of dog leashes in many graves from Birka, Tuna/Alsike, and Tuna/Badelunda suggests that dogs equipped with leashes may have possessed particular value or fulfilled specific functions, for example in relation to hunting or as markers of their owner’s status. Leashes, especially iron chains or richly decorated bronze junction pieces used in leather leashes, may not always have been deposited in graves, possibly due to their cost or restricted use.
A chronological analysis indicates a discernible development in dog leash forms. During the Vendel Period and the early Viking Age, particularly in Sweden, iron chain leashes appear to predominate in the archaeological record. In the ninth century, and at least into the late tenth century, such chains remain in use in Norway. At the same time, leather leashes incorporating cast bronze fittings appear in the ninth century. In Denmark, the first half of the tenth century sees the emergence of exceptionally elaborate junction pieces, likewise cast in bronze, intended for leather leashes, such as those from Ladby and Lille Lihme. The diversity of forms and materials suggests that dog leashes were neither uniform nor marginal elements of equipment, but objects of considerable value, both material and symbolic.
The analysis has identified a set of features characteristic of different types of dog collars, leashes, and swivels. At the same time, it has shown that swivels occur in large numbers and in the vast majority of cases should be regarded as belonging to equestrian equipment or other forms of harness. Consequently, not every swivel or iron chain can be automatically interpreted as an element of canine equipment. The same caution applies to fittings associated with dog collars. The definition of characteristic features of dog equipment may therefore provide a useful framework for the interpretation of future finds.
References
Arbman, H. 1943. Birka I. Die Gräber. Tafeln. Stockholm: Almqvist & Wiksell.
Arbman, H. 1943. Birka I. Die Gräber. Text. Stockholm: Almqvist & Wiksell.
Arwidsson, G., and G. Berg. 1983. The Mastermyr Find: A Viking Age Tool Chest from Gotland. Stockholm: Kungl. Vitterhets Historie och Antikvitets Akademien.
Blindheim, C. 1981. “Ryttermedaljongen fra Gokstad - en fremmed i vår vikingtidskunst.” In Gokstadfunnet: et 100-års minne = The Gokstad excavations: centenary of a Norwegian Viking find, edited by Einar Wexelsen, 30–31. Sandefjord: Sandefjordmuseene.
Blindheim, C., B. Heyerdahl-Larsen, and R. L. Tollnes. 1981. Kaupang-funnene. Bind I. Norske Oldfunn XI. Oslo: Universitetets Oldsaksamling.
Bonde, N., and A.E. Christensen. 1993. “Dendrochronological Dating of the Viking Age Ship Burials at Oseberg, Gokstad and Tune, Norway.” Antiquity 67 (256): 575-583.
Brøgger, A. W., H. Shetelig, and G. Gustafson, eds. 1926. Osebergfundet. Vol. I. Kristiania: H. Aschehoug & Co. (W. Nygaard).
Brøgger, A. W., H. Shetelig, and G. Gustafson, eds. 1928. Osebergfundet. Vol. II. Kristiania: H. Aschehoug & Co. (W. Nygaard).
Christensen, A. E., A. S. Ingstad, and B. Myhre, eds. 1992. Osebergdronningens grav: vår arkeologiske nasjonalskatt i nytt lys. Oslo: Schibsted.
Gansum, T. 2018. “The Royal Viking Age Ship Grave from Gokstad in Vestfold, Eastern Norway, and Its Link to Falconry.” In Raptor and Human – Falconry and Bird Symbolism throughout the Millennia on a Global Scale, edited by K.-H. Gersmann and O. Grimm, 717–724. Advanced Studies on the Archaeology and History of Hunting 1.1–1.4. Kiel/Hamburg: Wachholtz Verlag.
Grieg, S. 1928. “Jernlenker og hundelenker.” In A. W. Brøgger and H. Shetelig (eds.), Osebergfundet, vol. II, 248–249. Kristiania: H. Aschehoug & Co. (W. Nygaard).
Grieg, S. 1947. Gjermundbufunnet. Norske Oldfunn VIII. Oslo: Universitetets Oldsaksamling.
Gustafson, A. 1904a. Dagbok fra utgravningene ved Oseberg I. Manuscript, Kulturhistorisk museum, Universitetet i Oslo.
Hinsch, E. 1958. “Gokstadhøvdingens jaktransel?” Viking. Tidsskrift for norrøn arkeologi 21/22: 175–201.
Lindblom, C., and C. Adamsen. 2014. “Gemmer Jelling-egnen på ukendte vikingeskatte?” Videnskab.dk – Kultur & Samfund.
Magnell, O., E. Iregren, and K. Jennbert. 2023. “Breeding, Feeding and Mobility among Dogs in the Elite in the Merovingian Period – Analysis of Dogs in the Boat Grave Nabberör, Öland, Sweden.” Preprint, SSRN.
Müller-Wille, M. 1978a. „Frühmittelalterliche Prunkgräber im südlichen Skandinavien.” Bonner Jahrbücher 178: 633–652.
Myhre, B. 2015. Før Viken ble Norge: Borregravfeltet som religiøs og politisk arena. Norske Oldfunn XXXI. Stavanger: Arkeologisk Museum, Universitetet i Stavanger.
Nicolaysen, N. 1882. Langskibet fra Gokstad ved Sandefjord. Kristiania
Nordahl, E. 2018. Valsgärde 14. Acta Sepulcreti Valsgaerdiae Regiae Universitatis Upsaliensis I. Uppsala.
Öhman, I. 1983. The Merovingian dogs from the boat-graves at Vendel. In: J. P. Lamm and H. Ä. Nordström (eds.). Vendel Period Studies. Transactions of the Boat-Grave Symposium in Stockholm, February 2–3, 1981. Stockholm, 167–182.
Petersen, J. 1919. De Norske Vikingesverd: En typologisk-kronologisk studie over vikingetidens vaaben. Kristiania: J. Dybwad
Petersen, J. 1951. Vikingtidens redskaper. Oslo: J. Dybwad.
Rygh, O. 1885. Norske oldsager: ordnede og forklarede. Christiania: Alb. Cammermeyer.
Sørensen, Anne C. 2001. Ladby: A Danish Ship-Grave from the Viking Age. Copenhagen: National Museum of Denmark.
Stylegar, Frans-Arne H., and Ragnar L. Børsheim. 2021. “Gjermundbufunnet – en småkonges grav med østlig tilsnitt på Ringerike.” Viking. Norsk arkeologisk årbok LXXXV: 89–122.
Thorvildsen, K. 1957. Ladby-skibet. Nordiske Fortidsminder, Serie VI, Bind 1. København: Det Kongelige Nordiske Oldskriftselskab.
Harald Hrodberhtsson 2026