iron 'slave-collars' from lagore crannog, co. meath
TRANSCRIPT
Royal Irish Academy is collaborating with JSTOR to digitize, preserve and extend access to Proceedings of the Royal Irish Academy. Section C: Archaeology, Celtic Studies, History, Linguistics, Literature.
http://www.jstor.org
Iron 'Slave-Collars' from Lagore Crannog, Co. Meath Author(s): B. G. Scott Source: Proceedings of the Royal Irish Academy. Section C: Archaeology, Celtic Studies, History,
Linguistics, Literature, Vol. 78 (1978), pp. 213-230Published by: Royal Irish AcademyStable URL: http://www.jstor.org/stable/25506356Accessed: 29-06-2015 16:40 UTC
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/ info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].
This content downloaded from 5.68.165.228 on Mon, 29 Jun 2015 16:40:56 UTCAll use subject to JSTOR Terms and Conditions
[213]
8.
IRON ' SLAVE-COLLARS
' FROM LAGORE CRANNOG, CO. MEATH
By B. G. Scott
Queen's University /Ulster Museum, Joint Conservation Laboratory, Belfast.
(Communicated by J. Raftery, m.r.i.a.)
[Received, 23 June 1977. Read, 9 October 1978. Published, 1 December 1978.J
Abstract
The fine two-piece collar recovered from the lowest occupation at Lagore crannog, Co. Meath, provides much interesting insight into native Irish smith work, and its relationship to that practised in Britain. The long chain attached to the collar is of a style readily recognisable in late Roman Britain, while the inspiration for the collar decoration is suggested as originating in
Anglo-Saxon workmanship. Technical examination suggests that the collar as it now stands consists of two distinct halves made at different times and by different smiths.
The plain one-piece collar, easily paralleled from later pre-Roman Iron Age times through to the late Roman period in Britain, is also discussed and, along
with the two-piece collar, provides evidence for the reconstruction of a third, previously unsuspected collar now represented only by a fragment.
Consideration is given to the question of how the techniques behind the manufacture of these pieces reached Ireland, and of the more general question of the transmission of technical ideas in first millenium A.D. Ireland.
1. Introduction
Among the many fine objects recovered from the royal crannog of Lagore, Co. Meath, and now in the collections of the National Museum of Ireland, are a
superb two-piece decorated iron collar with its attachment-chain still in place, a fragment of a second two-piece collar, a one-piece collar closely parallel to
collars originating in the British pre-Roman Iron Age, and three short lengths of chain. Of these objects only the decorated collar came from the excavation
of the site carried out by H. O'N. Hencken; the rest were recovered by antiquar ians at different times between the date of the discovery of the crannog in the
early nineteenth century and the excavations from 1934 to 1936.
The decorated collar was sealed by the lowest level of occupation. Hencken
suggested a date around the mid seventh century A.D. for the start of building activity at the site, a date based both on his interpretation of archaeological association and on an interpretation by Liam Price of the literary evidence which seems to suggest that the construction of the royal crannog commenced about
651 A.D. Although Hencken's dating has been the subject of discussion and
criticism among Irish archaeologists, no one has, as yet, published alternative
dating based on a detailed reassessment of the site assemblage. There seems
to be no reason to alter significantly the seventh-century date for Period I at
PROC. R.I.A., VOL. 78 SECT. C [H8]
This content downloaded from 5.68.165.228 on Mon, 29 Jun 2015 16:40:56 UTCAll use subject to JSTOR Terms and Conditions
214 Proceedings of the Royal Irish Academy
Lagore on the basis of any evidence provided by the detailed examination of the
collars and chains described below, since it will be argued that manufacture of
the decorated piece dates to no later than the start of the seventh century A.D.
This paper is a re-evaluation of the significance of the two collars, collar
fragments, and the associated chains, based on a study of the way in which
they were made. From this it can be concluded that the decorated two-piece collar is made up of two unmatched halves representing two different collars of similar form, and that the manner of construction shows a strong foreign influence on its workmanship. It is also suggested that the collar-fragment
represents a new type which is intermediate in conception and construction
between the decorated two-piece collar and the plain one-piece Iron Age type. The origin of the technique of making the chains is discussed, and there is also a
consideration of the mechanism of transmission of technical knowledge and ideas
into and from Ireland.
2. The Decorated Two-piece Collar from the 1934-6 Excavations
(Pis. I and II)
(a) Description
The complete collar as we now have it is made up of two semi-circular halves
secured at the front by an arrangement of large links (PL la) and fastened at
the back by a riveted hinge (PL lb). The outer curved surfaces of the two halves are decorated by the application of strips of twisted square-sectioned bar (PL II) and the chain attachment is incorporated into the decoration of one side. For
the sake of convenience the half of the collar with the chain attachment will be
referred to as side 1, the other as side 2.
On side 1 (PL Ha) the decoration is made up of four twisted square-sectioned bars arranged in two pairs on either side of a plain central strip, also square in
cross-section. The plain strip is raised into a loop in one place to make the
attachment point for the chain. On either side, the pairs of twisted bars are
arranged so that their twists oppose to produce a pleasing herringbone pattern. Side 2 also has two pairs of twisted bars on either side of a plain central strip.
Here, however, the twists all run in the same direction so that no herringbone effect is produced (PL lib).
Side 1 is noticeably larger and heavier than side 2, forming the arc of a circle
whose inner radius is 7-3cm, outer radius 8-9cm. The maximum width of the
inner surface of side 1 is 3*6cm. Side 2 forms the arc of a circle with inner radius
8-2cm, outer radius 8-9cm: the maximum width of the inner surface is 2-8cm.
The chain is of interest both in the way in which its finks are formed and
joined together and also in that they taper in size from the attachment at the
collar to the present end of the chain. The biggest links are found just above
the point of attachment of the chain, where it passes through the arrangement of links forming the front closure mechanism. They have been forged from a
stock whose cross-section dimensions are 0'6 by 045cm. The link size here is
3*8cm in length by 2-3cm maximum width. At the other end of the chain the
This content downloaded from 5.68.165.228 on Mon, 29 Jun 2015 16:40:56 UTCAll use subject to JSTOR Terms and Conditions
Scott?Iron '
slave-collars '
from Lagore Crannog 215
smallest links occur; they measure 3-6cm long by 2*lcm in maximum width.
In contrast to the large links, these have been made from a thin iron wire with a cross-section of 02cm by 0-2cm. Although the complete chain is 3-65m long, the decrease in size of the links suggests that only the first two metres could
really be expected to stand up to the determined efforts of a would-be escapee, with the remaining links seeming unreliable as a means of securing the wearer
be it man or animal.
(b) Manufacture (Fig. 1)
Fig. 1 shows the various components (not to scale) and stages involved in the
making of the collar.1 Initially the smith gathered together two large billets of
iron2 and ten pieces of thin square-sectioned bar, nine of the same length, the
tenth somewhat longer. Eight bars of the same length were twisted, leaving a
short length at each end in the original condition. The long bar was bent in the middle to produce a loop and also effectively to reduce its length to that of the
others (Fig. la), while the tenth bar was left unaltered. The two large billets were drawn down to flat rectangular plates of flattened trapezoidal section.
Rectangular ' bites
' were taken from the corners of each plate leaving them
with an independent flap of thinned-down metal along each edge (Fig. lb).
The bars were divided into two groups, each consisting of twyo pairs of twisted bars and one plain bar. One group was arranged so that the twist of each bar ran in the opposite direction to that of its neighbours, and the looped strip was
laid between the two pairs. The other group was arranged in two pairs on either
side of a plain strip, but with the twists all running in the same direction. The
ends of the strips in each group were welded together to produce two composite
panels3 (Fig. lc).
Two small rectangular pieces were cut out of each side of one end-flap on
one of the plates, leaving a roughly square ' tail
' connected to the plate by a
thin strip of metal (Fig. Id). A single rectangular hole was cut in one of the
end-flaps of the other plate (Fig. le). Both strips were then bent to the required curvature.
The decorative panel with the loop attachment and herringbone patterning was placed against the plate with the rectangular perforation in the end-flap, and
this flap was bent round a piece of thin bar of circular section and over the welded end of the panel. The end-flap was then welded to the upper face of the end of the panel to cover the plain ends and to hold it in place, and the panel was bent
round to lie flush against the curved surface of the plate(Fig. If).
1 For the sake of simplicity, the manufacture of the complete collar is described as if carried out in one operation and by the one smith, although this was almost certainly not the case.
2 The material used in the manufacture of all the collars is ferrous metal, although without
metallographic examination, it is impossible to specify iron or steel, or a mixture. 3 One of the outside strips of side 1 has its plain end too wide to fit into the otherwise
rectangular outline of the panel. To overcome this, the smith has made it wedge-shaped to make it fit (see PI. lia).
This content downloaded from 5.68.165.228 on Mon, 29 Jun 2015 16:40:56 UTCAll use subject to JSTOR Terms and Conditions
2l6 Proceedings of the Royal Irish Academy
Fig. 1?Stages in the manufacture of the decorated two-piece collar. (Not to scale)
This content downloaded from 5.68.165.228 on Mon, 29 Jun 2015 16:40:56 UTCAll use subject to JSTOR Terms and Conditions
Scott?Iron '
slave-collars '
from Lagore Crannog 217
The second end-flap was bent round another, thicker, bar of round section
until the end was almost flush with the plain ends of the composite panel. Before it was welded into place, a large link with a slight waist was inserted into the
loop formed and the end-flap was then welded down over the end of the panel. The same procedure was carried out with the other composite panel and plate
(Fig. lg) to give side 1 with two loops projecting laterally at one end, and side 2
with a single, centrally-placed, projecting loop. The single loop of side 2 was
then inserted between the two loops of side 1, and a short rod of circular section
passed through all three. The ends of this bar were hammered into rivet heads, thus forming the permanent hinge (see PL lb). The large finks in the other ends
of the two sides form, with the chain, the front closing mechanism (Fig. lh). The link in side 2 is free to move in three directions and can be manipulated over the fink of side 1. If, in this position, the chain is passed through the end
link of side 1, the collar is effectively secured. The waist in the link of side 1
restricts the movement of the chain and presses the link of side 2 across the gap between the two sides of the collar. This, in turn, prevents the link of side 1 from
being manipulated sufficiently far to allow the chain to be slackened and the two halves of the collar pushed apart. As well as the actual closing mechanism, it is important to note that the chain, which forms an integral part of it, comes
from outside the immediate area of the mechanism (i.e. from the side of the
collar). We shall return to this point later.
At some stage, the projecting end-loop of side 2 was broken off and replaced
by a ' strap-end
' device (similar in shape to that labelled
' 2
' in PL IV) to
remake the hinge-joint. This repair has not been carried out in a particularly satisfactory fashion, as the ends of the outer-most strips of the decorated panel are left exposed (PL I). As pointed out above, the function of this flap, was to
hold the composite panel in place, and in doing so, to hide the plain ends of the
strips, and, in addition, to form the hinge.
3. The Chains and Their Fabrication (PL III and Fig. 2) The chain attached to the decorated collar (PL la) and the three fragments
(PL III) all have a link type extremely rare in Ireland. Instead of the usual
rectangular, oval or figure-of-eight links, the Lagore chains consist of links whose shape is that of a doubled-over figure-of-eight. This form, ultimately Roman in origin, is presumably a development of the figure-of-eight fink, since
this is an intermediate stage in its production. Figure 2 is a diagram (not to
scale) showing how the links were formed and joined together. After the stock had been forged into a roughly rectangular shape and the two ends scarf-welded
(Fig. 2a), a waist was formed by hammering towards each other the two long sides of the rectangle (Fig. 2b). The link was then bent over in the middle until
the two loops at each end touched at their tips (Fig. 2c). Although the hammering of the two loops and the waist together might accomplish a slight increase in
strength, the main function of this was probably to cut down the degree to which
the chain could tangle. It is noticeable that when handling the long chain
attached to the decorated collar where the tips of the loops of each link just touch, there is a strong tendency for the finks to get caught up with each other.
This content downloaded from 5.68.165.228 on Mon, 29 Jun 2015 16:40:56 UTCAll use subject to JSTOR Terms and Conditions
2i8 Proceedings of the Royal Irish Academy
Fig. 2?Stages in the manufacture of the chain-links. (Not to scale)
One interesting feature of chains made up in this way, which sets them apart from those of Llyn Cerrig Bach and Standfordbury types [infra) is that they can
only easily be lengthened by one link at a time. An unbent fink is passed through the loops of the link before it and bent over into the same shape (Fig. 2d). With
these chains, a new link is added by passing it, unbent, through the end-loops of the previous link, and not the loop formed when the link was bent double.
On Pis. lb and Ha it can be seen that it is these two end-loops which join up with the loop incorporated into the composite panel of side 1. While this
arrangement then points the link in the correct direction for further finks to be
added easily, it must mean that unless the smith involved himself in a much more
difficult and complicated piece of forging than was necessary, the first link must
have been fixed to the attachment point before the strip was incorporated into
the composite panel. One advantage of this technique is that the number of
individual finks required can be forged to the unbent, waisted stage (Fig. 2b) before the chain is actually put together. Assuming that we had two lengths of
chain made up of this type of link, and with the links all pointing in the same
direction, then to join them together would require that the ends of the joining fink be left unwelded until the basic shape had been formed and the link inserted
through the two loops of the link at one end of one of the chains?a complex process. Chains with figure-of-eight links, which are joined together with the
short axis of symmetry at 90? to those of the links on either side (e.g. Fox, 1946,
38, fig. 20) could readily be lengthened at either end or joined together. They are also less liable to tangle.
4. The One-piece Plain Collar (PL IV and Fig. 3) This piece can be paralleled exactly among collars which first appear in the
pre-Roman Iron Age of Britain (e.g. at Llyn Cerrig Bach?Fox 1946), and con
tinue on until at least Late Roman times (e.g. the fourth century A.D. Great
Chesterford find?Neville 1856). It is made from a bar of trapezoidal cross
section. The ends have been wedged out, folded over and welded back on
themselves to give two end-loops of different shapes. The larger of these has two
dissimilar large links through it, while the smaller holds one link only. These
three large links, along with the chain which would originally have been attached
to one of the pair of links in the same loop, form the closing mechanism shown
This content downloaded from 5.68.165.228 on Mon, 29 Jun 2015 16:40:56 UTCAll use subject to JSTOR Terms and Conditions
Scott?Iron * slave-collars
' from Lagore Crannog 219
Fig. 3?Reconstruction of the closure mechanism of the plain one-piece collar. (Not to scale)
This content downloaded from 5.68.165.228 on Mon, 29 Jun 2015 16:40:56 UTCAll use subject to JSTOR Terms and Conditions
220 Proceedings of the Royal Irish Academy
in Fig. 3. The main differences between this fastening mechanism and that of
the decorated two-piece collar [supra, section 2b) is that the chain needs to be
passed through both free links and also that it commences in the immediate area of the closing mechanism. If the chain were not passed through both free
links, it would be possible, if only one link were used, to manipulate the chain so that the collar could be opened.
5. A 4 New
* Collar from Lagore (PL V and Fig. 4)
In PL III, one of the lengths of chain described above can be seen to have
a ' strap-end
' device joined to it by a large oval link (the piece is marked
' 2
'
in Pis. Ill and V). Examination of this strap-end shows it to consist of two
folded-over plates connected by a thin strip of metal raised into a loop through which the large oval link is passed. One plate has a ragged end that is the result
of an ancient break, while the other exhibits no such signs of damage. Between
two plates are sandwiched the broken ends of five bars of rectangular section
(PL V). The edges of the two plates have been hammered together tightly and
it is therefore impossible to tell whether the ends of these strips have been
welded together or not, although this is perhaps suggested by the position in
which four of the strips lie. One puzzling feature about the original publication of this piece is that although the fragment of chain with the large oval link was
drawn for a paper by Talbot4 (Talbot 1849, fig. opp. p. 104), the ' strap-end
' is
not shown. This is odd since there is no way in which it could have been added
recently without the joint being detectable. It must, however, be pointed out
4 I am grateful to U. O Meadhra for drawing my attention to this illustration.
This content downloaded from 5.68.165.228 on Mon, 29 Jun 2015 16:40:56 UTCAll use subject to JSTOR Terms and Conditions
Scott?Iron '
slave-collars '
from Lagore Crannog 221
that if the other artifacts shown in the same figure are those redrawn for Hencken's excavation publication, we can assume that Talbot's drawings are not accurate.
The only reasonable explanation of this piece is that it represents one end of a collar basically of similar form to the two-piece decorated example, and totally different from the Iron Age type, but owing something to each for its construction.
Figure 4 is a tentative reconstruction of what this collar might have looked
like, based on observation of the surviving fragment and of the other two collars described above. The illustration shows a collar made up of two halves, each formed from strips of rectangular-sectioned bar welded together. There are ' strap-end
' attachments at each end which help to hold the strips together and
also to form loops for the rivet of the hinge and links of the front closing mechan ism. The hinge is the same as on the two-piece decorated collar, while the closing
mechanism is similar to that of the specimen of Iron Age type.
The arguments for this reconstruction are as follows :
(i) The surviving end of one of the halves of the collar is attached to the chain
by a large link similar to the chain attachment of the Iron Age collar. However, unlike the Iron Age collar, there is room for one link only to pass through the
end-loop and it is not unreasonable to suppose that the end-loop of the other
half of the collar would have been of similar proportions.
(ii) The fact that the break in this piece is clean and not the result of corrosion
during its period of deposition suggests that there was a weakness here. If the
composite panels had been mounted on backing plates (as on the two-piece decorated example) it could be expected that this would reinforce them sufficiently to prevent the fracture occurring where it did.
(iii) The weakness suggested would have been the result of placing the chain
attachment at one end rather than in the middle of the collar with support from a solid backing plate. As only one link could be inserted through the
surviving end-loop the chain could be pulled backwards and forwards by straining
against the point of attachment of the free end. If the collar had been made in
one piece we would expect the back region to take the full fatiguing stress and
fracture to occur here. Unless more has broken off the fragment since it was
discarded (which seems unlikely), it will be obvious that if there had been a back
hinge to allow free movement the stress would have been concentrated at the
point of chain attachment. It is thus interesting to find that the rectangular
strips have sheared off at exactly the point where the reinforcing cover of the tails
of the strap-end stop (PI. V).
6. Discussion
For the sake of simplicity, the manufacture of the decorated collar has been
described [supra section 2b) as if carried out by one smith. Apart from the
noticeable difference in the size of the two sides of this collar, there is also a
marked difference in the quality of workmanship displayed by each. Side 1 is
by far the better made. The herringbone pattern is made up of strips which show
[H8a]
This content downloaded from 5.68.165.228 on Mon, 29 Jun 2015 16:40:56 UTCAll use subject to JSTOR Terms and Conditions
222 Proceedings of the Royal Irish Academy
a uniformity of torsion along their length and also when compared to each other.
The torsion of the strips of side 2 is uneven in comparison with those of side 1.
This seems to suggest that the two halves were not made by the same smith and
also that the smith who made side 1 was a more competent craftsman than the
maker of side 2.
The most likely explanation is that side 1 represents an original collar in
which both sides were matched in pattern, size, and workmanship, and that
side 2 represents an inferior copy of the original half of the collar made by a less
competent smith who did not elect to use (or did not grasp the significance of)
opposed twists for decoration. A good reason for the joining together of sides 1
and 2 would be that side 2 was a substitute for an original too badly damaged for repair. Side 1 represents an extremely fine piece, and there is no difficulty in imagining that the owner would rather have had it repaired than discard it.
Now, either side 2 was specially made to fit side 1, or it was cannibalised from a
complete but inferior copy of the original. In the light of the discovery of the
fragment of a collar of related type, and because it seems unlikely that a smith
deliberately making a replacement side could have got both dimensions and size
wrong, the balance of evidence is for the existence of a ' second-hand
' copy,
represented by side 2, of the original represented by side 1. This also helps to
throw light on the cause of the repaired break on side 2 (supra, section 2b).
We have seen in section 5 that one of the main points of weakness on a collar
of this type would be the point of chain attachment and that the use of a backing
plate would add mechanical strength. Yet on side 2 we see that the whole
end-flap at the hinge has been broken off. It would seem more likely that, if a
fracture were to have occurred naturally here, the ends of the strips of the
composite panel would have sheared also. It is also likely that the amount of
stress needed to accomplish this would have sheared the rivet first, unless
breakage were deliberate. To unhook the two halves of the collar supplying the
replacement, it would have been necessary only to cut the rivet of the hinge. It is probable therefore that the end-flap of side 2 was deliberately removed
because the projecting end-loop was not of the right shape and-or size to fit
between the projecting loops of side 1. It has already been pointed out that the
repair is somewhat of a botched job because it fails to cover properly, and thus
secure, the outermost strips of the composite panel. For a smith to copy the
method of construction of the collar represented by side 1, he must at least have
understood the importance of using the folded-over-end and side-flaps to secure
the composite panel in place. This would suggest that the smith who repaired side 2 was not the smith who made it. A possible candidate is the smith who
made the collar described in section 5, because of the similarity between the
repair of side 2 and the surviving terminal of the new collar. He did not use a
backing plate, but simply secured the strips together by welding them into
a strap-end. To make the end-loop of size 2 fit it would probably have seemed
quickest to employ the device that he had used previously. However, it is still
possible that we are dealing with three collars made by three separate smiths,
working at different times.
This content downloaded from 5.68.165.228 on Mon, 29 Jun 2015 16:40:56 UTCAll use subject to JSTOR Terms and Conditions
Scott?Iron '
slave-collars f
from Lagore Crannog 223
With the exception of a fine gold chain found in 1842 at Newgrange, with
other gold ornaments (the Conyngham find?Carson and O'Kelly 1977, 53-54, and PL III), it is most difficult to find parallels in Ireland for either the collars or chains from Lagore. The chain from among the Conyngham gold ornamhnts consists of fifty links of bent-over figure-of-eight form, and was found with two
rings and two bracelets with strong Romano-British affinities dated to the period second to fifth centuries A.D. The only other Irish examples of the link type known
to the writer come from the three-link chain attached toa mount (described as a ' chariot fitting ') from a find at Navan, Co. Meath (Wood-Martin 1895,248, fig. 50A),
and a three-link chain from Ballinderry crannog no. 1 (Hencken 1936, 146, fig. 5). The Navan piece can probably be ascribed to the ninth or tenth centuries, while
the chain from Ballinderry comes from a level suggested by the excavator to
date to the eleventh century onwards.
In Britain, however, the links of the Lagore chains are identifiable as Piggott's Great Chest erf ord type, being identical, for example, to those seen on the remains
of cauldron suspension rigs from the Great Chest erf ord, Essex (Neville 1856, PL 3 and 4ff.) and Blackburn Mills, Berwickshire (Piggott 1955, 42-44, and
43, fig. 11, B-17: cf. Manning 1972, 233 note 63) hoards of ironwork, and
these may further reinforce the dating of the Newgrange find. We may also note the form in bronze in chain fragments from pit 20 at Richborough, Kent
apparently deposited some time after the late Claudian period (Bushe-Fox 1928, 30, and PL XV, 10). A significant parallel is provided by the links of the Sutton
Hoo cauldron suspension gear (Bruce-Mitford 1973, 38, fig. 16 and PL 23). Later examples include two bronze ch?telaines from Queen Victoria Street, London (Wheeler 1935, 189-190, and Pis. XVIII and XIX) and three unassoci
ated fragments also from London (Guildhall Museum Catalogue, PL XX, nos. 3, 4 and 7, and pp. 58-59). What may be a badly corroded link of this type occurs among the iron finds from Manching, in Bavaria (Jacobi 1974, 226, and Taf. 58, no. 877). An interesting mediaeval foreign parallel is provided by the silver chain links from grave 11 at Wartmannstetten in Austria (Hampl 1961, 26-29, 28, abb. 14). We may disregard Hencken s suggestion (1951,115-117) that the Lagore chains are to be classed with those from Bigbury Camp in Kent
(Boyd-Dawkins 1902, 211 and PL III), since the latter fall into Piggott's Stand
fordbury type (Piggott 1955, 26). We do encounter serious problems when looking for parallels for the original
decorated Lagore collar. One of a group of five one-piece collars joined together
by a padlock in the Great Chesterford hoard is made of a flat strip of iron with a raised cord decoration running down the centre (Neville 1856, 9 and PL 2.24). However, this decoration lies on the inside, and is probably not more than an
interesting example of relief forging rather than a parallel to the welding on of
twisted strips to a backing plate.5 The technique of opposing twists on separate rods to produce a herringbone pattern is fairly well known in later prehistory,
being seen in Ireland, for example, in the gold necklace from Duhallow, Co. Cork
(Windele 1861, 28f.) and the composite Broighter, Co. Deny torque. Individual
5 I am indebted to the referee for drawing this piece to my attention.
This content downloaded from 5.68.165.228 on Mon, 29 Jun 2015 16:40:56 UTCAll use subject to JSTOR Terms and Conditions
224 Proceedings of the Royal Irish Academy
twisted rods are quite common as bucket handles and the like both in Ireland
(as at Lagore?Hencken 1951, 114, fig. 47) and Britain (e.g. the Blackburn Mill
hoard?Piggott 1955, 43, fig. 11, B-12), and were also used as suspension rods in
cauldron chain complexes (e.g. the pieces in the Eckford, Roxburghshire hoard?
Piggott 1955, 25, fig. 5, E6-E9). But the placing of twisted rods together to make
composite herringbone-patterned panels seems to occur first with the six
examples incorporated into the Sutton Hoo cauldron suspension gear. It must
be noted that this complex has its best British parallel in the suspension gear from Great Chesterford, Essex (Neville 1856, PL 3, 4ff.; Piggott 1955, PL I),
where the beautifully wrought centre-piece (as well as the links) incorporating a
perfect reef-knot in iron, is very close indeed to the form of the Sutton Hoo
centre-piece. The Blackburn Mill centre-piece (Piggott 1955, 43, fig. 11, B-17) is similar, but less elaborate, lacking particularly the reef-knot.
The dating of the Sutton Hoo suspension gear is of importance to this study, on account of the herringbone panels. The Blackburn Mills hoard contains
strong Roman elements (some tools and the bronze patera?Piggott 1955, 9f.), and Piggott would see it as dating to some time in the first to second centuries, A.D.
It is accepted that the Great Chesterford hoard dates to the fourth century A.D. (Piggott 1955, 8; Manning 1972, 235-236). Since it is evident from the
repairs to the bronze cauldron from Sutton Hoo that this piece had been in use
for some time, and considering the ' heirloom
' nature of objects like the
Anastasius dish from the same deposit, it is not impossible that the Sutton Hoo
suspension gear could have been made roughly contemporary with the Great
Chesterford example, without complications over the mid seventh-century terminus ante quern now
placed on Sutton Hoo on numismatic grounds.
Examples, in iron, of the use of opposed twists to produce composite artifacts
bearing panels of herringbone decoration (not pattern-welded) are hard to find.
And the best parallel, as we shall see, for the decoration of side 1 of the Lagore collar may well come from the composite panels included in the Sutton Hoo
suspension gear. But the idea has a much wider application as the basis for the
herringbone decoration of a large number of pattern-welded weapons both from
Britain and abroad. And we may note that X-radiography of the sword blade
from the Sutton Hoo deposit revealed that the blade had a herringbone pattern. This is the simplest form of the technique (and the earliest ?) being seen, for
example, on swords from the great Nydam deposit (e.g. Wilson 1970, 38, fig. 17 ; Sch?rmann 1959). The technique was well known among the people who
assembled the Sutton Hoo grave goods (and chain suspension rigs for cauldrons are known from Scandinavia?Bruce-Mitford 1973, 39, and 91, note 31 : cf. Jacobi 1974, 112f? 114, abb. 28 and Taf. 34 no. 592 for other comparanda) and to
the writer at least, the derivation of the idea for the Sutton Hoo herringbone
panels from herringbone pattern-welding seems a likely explanation.
In both Sutton Hoo and Lagore examples, we see the combination of chains
with a highly distinctive link form with the basic element of herringbone pattern
welding. There seems little doubt that the Sutton Hoo chain links follow a
tradition well established in Roman Britain, and that the complex as a whole
This content downloaded from 5.68.165.228 on Mon, 29 Jun 2015 16:40:56 UTCAll use subject to JSTOR Terms and Conditions
Scott?Iron ' slave-collars
' from Lagore Crannog 225
draws its decorative inspiration both from Romano-British (i.e. the chains and
centre-piece with reef-knot) and Germanic (the herringbone panels) smithing. There is good archaeological evidence for contacts between Roman Britain and
Ireland during the currency in Britain of the Great Chesterford link type (see most recently Warner 1976), as well shown by the Newgrange find, although these are unlikely to have introduced the idea of pattern-welding. Indeed,
although contacts with the Germanic world which could have introduced
pattern-welded weapons to Ireland are also attested in the archaeological record (e.g. Jope 1955, Scott 1977), the first certain examples of pattern
welded artifacts are firmly of Viking origin. Of these, only three have a
herringbone pattern; one (a sword from a burial at Larne, Co. Antrim?
Fanning 1970) is dated to the tenth century.6 But it is surely too great a
coincidence that these two elements should combine independently in East
Anglia and Co. Westmeath, and then be found on pieces deposited at roughly the same time. Given that the chain is an integral part of side 1, that the
Newgrange chain is of a date towards the mid first millenium A.D. and that
the few other chains of the type are dated much later, also that herringbone
pattern is a later first millenium A.D. intrusive technique, then either side 1
of the Lagore (and its chain) is an import, or was it made by an Irishman
working directly under the influence of foreign smithing traditions, and we must
surely look to Britain, most likely to Anglo-Saxon workmanship of the sixth to
seventh centuries, as the source.
That a copy (represented by side 2) could be made of the original collar
(represented by side 1) in which the essential feature of the original decoration
is lost argues for workmanship, probably of a native smith, uninfluenced by the
smithing traditions of the first. The third collar, represented by the fragment, shows an imperfect understanding of the rationales behind the constructions
either of the imperfect copy or the original, and was probably made by an Irish
smith. The identity of the smith who made the original collar and its chain, and
the way in which it reached Ireland is of great interest since it raises questions about the mechanisms of transmission and reception of technological ideas in
early Ireland.
As well as strong evidence for close contact with Scandinavia in general, and
Sweden (where pattern-welding was well known?Th?lin 1967), there is also
evidence of an Irish connection in the Sutton Hoo assemblage. Henry (1965,
74-75) describes hanging-bowl escutcheons from the assemblage as *. . . obviously of Irish workmanship (my italics) ', and also sees a link between them and the
establishment of a monastic settlement by the Irish missionary, St. Fursa, at
nearby Cnobheresburgh around 630 A.D. on land granted by the Christian king
Sigeberht, a member of the ruling house of East Anglia, and presumably therefore
a relative of the person for whom the Sutton Hoo tumulus was erected. It is
historically attested that many Irish monastic communities, at least in their
6 The two other examples of herringbone patterning known to me from Ireland are a
' dagger
' from a burial at Eyrephort, Co. Galway (Raftery 1960) and a winged spearhead
simply localised to '
Ireland '
{in J. B0e Viking Antiquities in Ireland) both in the collections of the National Museum of Ireland.
This content downloaded from 5.68.165.228 on Mon, 29 Jun 2015 16:40:56 UTCAll use subject to JSTOR Terms and Conditions
226 Proceedings of the Royal Irish A cademy
earlier periods, numbered skilled metalworkers among their inmates, and that
metalworking in general was an integral part of their internal economic organisa tion ; indeed, remains of ironworking have been uncovered at Fursa's monastery
(Cramp 1976, 214).7 The magnificent garnet jewellery included in the Sutton
Hoo deposit strongly suggests that the Wuffingas had at their disposal the
services of a workshop under the direction of at least one master-craftsman.
And such a workshop need not have confined its activities to jewellery work, but could also have embraced fine smithing. A passage from the Laws of Ine of
Wessex composed c. 690 A.D. states that a smith was an essential part of a
nobleman's household (Schubert 1957, 62). We have similar evidence for master
craftsmen working for the Norwegian nobility from the Frostingslag (Capelle 1968, 93f.). Given the obviously good relations that existed between Fursa's
foundation and the East Anglian royal household, and the probable existence of
metal workshops in both, we have a potentially strong point of contact between
two schools of workmanship. Even if the chain complex from Sutton Hoo was
an heirloom, it seems quite possible that the ideas which it incorporates would
have been open to learning by Irish monastic smiths. And the links that were
maintained between the homeland and missionaries abroad provide one mechan
ism whereby the same combination of ideas could possibly have been transmitted
back to Ireland. The fact that Fursa could have made for this region to be welcomed
with such obvious warmth argues for pre-existing contacts that might have acted as transmitters of ideas before the formal establishment of the monastery.8
The copy of the original collar was made by a smith unfamiliar with the idea
of herringbone patterning, less skilful than the maker of the original and certainly unaware of the technique of producing herringbone pattern-welding. The reason
for making the copy is immaterial; what is of importance is that here we see one
example of foreign influence on native workmanship?the influence of form
unrelated to the underlying technical concept. The hybrid collar shows the way in which another smith borrowed ideas from two sources (the tradition of making
one-piece collars which stretches back to Iron Age times, and the two-piece type),
again on the basis of form unrelated to technology. The contrast between the purely mechanical copying of shape or the imitation
of a style whose underlying technology is imperfectly understood, and the
assimilation and adoption of new technical ideas presented by this interpretation is of importance since it is something encountered frequently in the archaeology of the first millennium A.D. in Ireland. A worker skilled in the reproduction of
shapes needs no more than a model from which to work. But if he is unacquainted
7 As well as the excavated evidence from Nendrum, Co. Down (Lawlor 1928), there are numerous references in the literature to ecclesiastical smiths. Of particular interest is a
story in Adomnan's Life of Columba (ed. and trans. Anderson and Anderson 1961, 392, 76b) which shows the presence of smiths in the lona community in the saint's lifetime. See also O'Neill's study (1967) of mining and metallurgy by monastic communities of the British Isles.
8 Were it to be demonstrated, for example from reassessment of imported pottery and other artifacts, that the building of the crannog commenced at a date significantly earlier than c. the mid 7th century A.D., this should not decrease the significance of the connections between the two pieces to any great degree. It would obviously lessen or remove the possibil ity that monastic contacts were responsible for the arrival of the original Lagore collar in
Ireland, thus necessitating a fresh interpretation of the transmission mechanisms.
This content downloaded from 5.68.165.228 on Mon, 29 Jun 2015 16:40:56 UTCAll use subject to JSTOR Terms and Conditions
Scott?Iron ' slave-collars
' from Lagore Crannog 227
with the raw material, or with a new technique for working a familiar material, he needs first-hand instruction to be able to exploit the innovation properly or at all.9
Two negative examples illustrating this are the apparent absence of either native Irish-made wheel-thrown pottery10 or the induced-draught iron smelting furnace (the shaft-furnace) in the first millennium A.D. For each, we have good evidence that the Irish were in contact with people to whom these techniques
were known and were, in the case of wheel-thrown pottery, actually importing the products (probably as the containers for perishables). If a new material, or
an old material used in a new way, is encountered to-day, a range of sophisticated laboratory study techniques can be brought to bear in establishing its identity.
On the other hand, unless the early worker could find out at first-hand how an
object was made in terms of manipulating the raw materials, he would have been faced with the most severe difficulties imposed by an essentially empiric tech
nology. Thus, while Irish potters might have tried to copy the shapes of imported wheel-thrown pottery, they would have been unable to reproduce the technique
by merely looking at the vessels. However, even if a small group of wheel-using
potters, or of iron smelters who used the shaft-furnace, had established workshops in Ireland where the natives of the region could come to observe and learn, there are factors in the organisation of production in early Ireland preventing these
techniques from spreading and being adopted and exploited.
Early Irish industry takes its ' organisation
' as a direct result of the ordering
of early Irish society and settlement (Scott 1973, 1978). In precisely the same
way that settlement is scattered and small-scale, so is early production. Here and
there, sites like Garannes, Ballycatteen, Clogher and Lagore produce industrial
debris that suggest a greater level of production than normal, but these are
exceptional. The concept of the ' town
' as a political, social, economic, and
industrial unit is totally foreign to Ireland before the establishment of permanent
Viking settlements. Even after this the concept was not adopted. Excavations
carried out in Viking Dublin (? R?ord?in 1971) reveal intensity and diversity of
industrial activity simply not encountered on native sites. Groups of professional artisans were operating workshops in the mainstream of European production
technologies, and making goods for trade as well as local consumption. With
native production on a level comparable with a nineteenth-century cottage
industry, opportunities for observation, assimilation, and dispersal of knowledge would have been infrequent. Furthermore, since the bulk of production seems to
have been directed towards immediate local consumption, the incentive for
8 An example of imperfect understanding of raw materials come from the forging of the iron Llyn Fawr Sickle, and the iron looped and socketed axes from Britain and Ireland
(Scott 1974b, 1 If.). Here, copies are made of existing bronze types in a new material whose
physical properties are less suitable to the cast shapes of the originals. 10 See Thesaurus Palaeohibernicus Vol. I, Milan Glosses on the Psalms, fol. 18b, gloss 4,
p. 23. A. roth cruind forsandenat nacer da innalestrai . . . nid crann f cr?ind immuambiat ocuan denum: i.e. a round wheel on which the potters make the vessels, or a round piece of wood about which they are while being made. This shows Irish monks acquainted
with the potters' wheel and, given familiarity with imported wheel-made wares in Ireland and contacts with expatriates, it should have been possible for people in Ireland to get an
idea of the technology, but not practical experience of it.
This content downloaded from 5.68.165.228 on Mon, 29 Jun 2015 16:40:56 UTCAll use subject to JSTOR Terms and Conditions
228 Proceedings of the Royal Irish Academy
experimentation and assimilation would be much less than in the context of an
urban industrial economy geared to trade. Travelling craftsmen might have
been vehicles for the spread of new technical ideas were it not for the absence of
evidence, literary or archaeological, for their existence. And our immigrant group of wheel-using potters and shaft-furnace smelters would thus only be able to
impart information which could not survive their passing.
Poor communications and the lack of opportunity or incentive to learn new
techniques may thus be cited as reasons for the very slow development of certain
production industries in first millennium A.D. Ireland. Although the rise of
monastic communities containing skilled artisans could have acted as industrial
foci, their influence would have been limited to that knowledge which could be
transmitted visually or orally?art-styles, religion, and literature. Otherwise, the factors outlined above would have restricted their role as disseminators of
practical, technical information. Yet another factor is the exceedingly small
number of genuine innovators, as compared to the very large number of workers,
in any industry.11 Where communications are good this should not prevent the
spread of new ideas,12 but where they are bad, new ideas may have* a very limited
range of travel. It is not surprising, therefore, that if we consider the arrival in
Ireland of the original decorated collar and its attached chain (and even the
nearby Newgrange chain) as an innovating influence, its effects are extremely limited. Even had the smith who made the original collar and chain come to
Ireland with his handiwork, we might not expect to see his influence have any
greater spread.
Conclusions
The dating of the deposition of the complete collar as it now stands must at
present be established by Hencken's interpretation of the evidence for the
period la of the site. Thus the original collar represented by side 1 can be no
later than the earlier seventh century. From the sequence of construction of the
collar, it should be assumed that the original arrived in Ireland some time before
this, to allow time for the copy to be made and for the two unmatched halves to
come together. It is unlikely that this would have occurred too close in time to
the arrival of the original, and manufacture of the copy.
A final point may be made about the function of the two-piece collars.
Although commonly referred to as ' slave collars ', it would seem odd that such
fine pieces would have been put to such lowly use. Byrne (1973, fig. 13, opp. p. 88) more reasonably suggests that the complete Lagore collar could have been used
to secure hostages. Another explanation is that they were not intended for men, but animals. A passage in the T?in B? C?ailgne which relates how the infant
11 In talking of innovators, we are referring to individuals or small groups who make
discoveries, as opposed to the people who possess a particular piece of information and transmit it to the unenlightened.
12 A good example of the spread of a new idea can be seen in the rise of beaten bronzework in barbarian Europe and the British Isles. It stems ultimately from the recognition of the
technique of annealing which allows bronze to be beaten out into very thin sheets which can then be used in the construction of various artifacts such as buckets, cauldrons and
body armour.
This content downloaded from 5.68.165.228 on Mon, 29 Jun 2015 16:40:56 UTCAll use subject to JSTOR Terms and Conditions
Scott?Iron '
slave-collars ' from Lagore Crannog 2,2,$
Setanta came to be called Cu Chullainn, tells how a fierce dog belonging to
Culann Cerd, and of which he was inordinately proud, came to be killed. In this
passage, the dog is described as being chained up because of its temperament
(O'Rahilly 1970, 24, 856-871). It is also interesting to note the occurrence at
Lydney Park, Gloucs.13, of a small bronze figurine of a dog (Wheeler 1932, 40,
88-9, and PL XXV), identified as an Irish wolfhound whelp, and which is depicted as wearing a large collar. Is it not possible that instead of being used to secure
slaves or hostages, the decorated Lagore collars might have been used for
favourite animals as a way of showing the esteem of the master for his pet (while at the same time keeping the
' pet
' safely restrained) ?
Acknowledgements
I should like to express my gratitude to the Director of the National Museum of Ireland, Dr. J. Raftery, for permission to study this group of objects, and for
his many helpful comments during the writing of this paper. Also to the National
Museum for providing the plates, to the Assistant Keeper of Irish Antiquities, Mr. M. FitzG. Ryan for much assistance, and to Mrs. D. Crone of the Department of Antiquities, Ulster Museum, who composed the illustrations. I am grateful, also, to the referee for his comments.
REFERENCES
Abbreviations
Arch. J. Archaeological Journal. J.G.A.H.S. Journal of the Galway Archaeological and Historical Society. J.R.S.A.I. Journal of the Royal Society of Antiquaries of Ireland.
Med. Arch. Mediaeval Archaeology. P.R.I.A. Proceedings of the Royal Irish Academy. P.S.A.S'. Proceedings of the Society of Antiquaries of Scotland.
U.J.A. Ulster Journal of Archaeology.
Anderson, A. O. and M. O., ed. and trans. 1961 Adomnan: Life of Columba. Oxford.
Boe, J. 1940 Viking Antiquities in Great Britain and Ireland, 3. Oslo.
Boyd-Dawkins, W. 1902 On Bigbury Camp. Arch. J. 59, 1902, 21 If.
Bushe-Fox, J. 1928 Second Report of the Excavation of the Roman Fort at Richbor
ough, Kent. Reports of the Research Committee of the Society of Antiquaries of London, no. VII, London.
Bruce-Mitford, R. L. S. 1973 Sutton Hoo. London.
Byrne, F. 1973 Irish Kings and High Kings. London.
Capelle, T. 1968 Das Metallschmuck von Haithabu. Neumunster.
Carson, R. A. G. and O'Kelly, C 1977 A catalogue of the Roman coins from New
grange Co. Meath and notes on the coins and related finds P.R.I.A. 77 C (2), 35-53,
Catalogue of the Collection of Antiquities in the Guildhall Museum (2nd edn.), London 1908.
Cramp, R. J. 1976 Monastic Sites. In ed. D. M. Wilson, The Archeology of Anglo-Saxon England. London, 201-252.
Fanning, T. 1970 Viking grave-goods from Larne. f.R.S.A.I. 100 (i), 71-78.
Fox, Sir C. 1946 A find of the Early Iron Age from Llyn Cerrig Bach, Anglesey. Nat. Mus. of Wales, Cardiff.
Hampl, F. 1961 Ein fr?hgeschichtliches Gr?berfeld in Wartmannstetten, pol. Bez.
Neunkirchen, N.?. Arch ologia Austr?aca, 29, 18-37,
Hencken, H. O'N. 1936 Ballinderry Crannog no. 1. P.R.I.A. 43 C (5), 103-239.
Hencken, H. O'N. 1951 Lagore crannog, Co. Westmeath: an Irish royal residence of the 7th-10th centuries. P.R.I.A. 53 C (1), 1-247.
13 I am grateful to R. B. Warner for drawing my attention to this piece.
This content downloaded from 5.68.165.228 on Mon, 29 Jun 2015 16:40:56 UTCAll use subject to JSTOR Terms and Conditions
230 Proceedings of the Royal Irish Academy
Henry, F. 1965 Irish Art in the Early Christian Period to A.D. 800. London.
Jacobi, G. 1974 Werkzeug und Ger?t aus dem oppidum von Manching, Wiesbaden.
Jope, E. M. 1955 A spearhead of Germanic type from Co. Fermanagh. U.J.A. 18, 16-17.
Lawlor, H. 1928 The monastery of St. Mochoi at Nendrum. Belfast.
Manning, W. H. 1972 Ironwork hoards in Iron Age and Roman Britain. Britannia 3, 224-250.
Neville, R. C. 1856 Description of a remarkable deposit of Roman antiquities of iron, discovered at Great Chesterford, Essex, in 1854. Arch. J. 13, 1-13.
O'Neill, H. 1967 Monastic mining and metallurgy. Metals and Materials 1 (6), 182-190.
O'Rahilly, C. ed. and trans. 1970 T?in B? C?alnge from the Book of Leinster. Dublin. ? Ri ordain, A. B. 1971 Excavations at High Street and Winetavern Street, Dublin,
Med. Arch. 15, 73-85.
Piggott, S. 1955 Three metalwork hoards of the Roman Period from Southern Scotland. P.S.A.S. 87, 1-50.
Raftery, J. 1960 A Viking Burial in County Galway. J.G.A.H.S. 29 (i, ii), 3-6.
Schubert, H. 1957 History of the British Iron and Steel Industry. London.
Sch?rmann, E. 1959 Untersuchungen an Nydam-Schwerten. Archiv f. Eisenh?ttenwesen 30 (3), 121-130.
Scott, B. G. 1973 Early Irish ' tribes
' and
' tribalism '. Ogam, XXII-XXV, 1970-1973,
in press. Scott, B. G. 1974a The application of techniques of physical examination to archaeolo
gical research. In Perspectives in Irish Archaeology, ed. B. G. Scott. Belfast, 107-120.
Scott, B. G. 1974b Notes on the transition from bronze to iron in Ireland. Irish Archceol. Res. Forum 1 (1), 9-21.
Scott, B. G. 1977 Metallographic examination of some early Irish iron tools and weapons. P.R.I.A. 77 C (12), 301-317.
Scott, B. G. 1978 The introductions of non-ferrous and ferrous technologies to Ireland: motives and mechanisms. In Fifth A tlantic Colloquium Proceedings, Dublin. Forthcoming.
Talbot, J. 1849 Memoir on some ancient arms and implements found at Lagore near
Dunshaughlin, County of Meath; .... Arch. J. 6, 101-109.
ThAlin, L. 1967 Metallografisk unders?kning av ett vendeltidska praktsv?rd. Fornv?nnen 62, 305-324.
Warner, R. B. 1976 Some observations on the context and importance of exotic material in Ireland, from the first century B.C. to the second century, A.D., P.R.I.A. 76 C (12), 267-292.
Wheeler, R. E. M. 1932 Excavation of the Prehistoric, Roman and Post-Roman site in
Lydney Park, Gloucs. Reports of the Research Committee of the Society of Antiquaries of London no. IX. London.
Wheeler, R. E. M. 1935 London and the Saxons. London.
Wilson, D. 1970 The Vikings and their Origins. London.
Windele, J. 1861 Ancient Irish Gold, U.J.A. 33, 28-50.
Wood-Martin, W. G. 1895 Pagan Ireland: An Archaeological Sketch. London,
This content downloaded from 5.68.165.228 on Mon, 29 Jun 2015 16:40:56 UTCAll use subject to JSTOR Terms and Conditions
a
Lagore crannog. Decorated two-piece collar, a: Front view showing closure mechanism, b: Back view showing hinged joint and repair to side 2.
Courtesy National Museum of Ireland.
Lagore crannog. Decorated two-piece collar, a: Side 1 showing herringbone pattern and chain attachment, b: Side 2.
Courtesy National Museum of Ireland. . .
This content downloaded from 5.68.165.228 on Mon, 29 Jun 2015 16:40:56 UTCAll use subject to JSTOR Terms and Conditions
PROC R.I.A., VOL. 78, SECT. C Plate III Plate IV
Lagore crannog. Chain fragments among odd finds. The modern join between a and b is marked with an arrow.
Courtesy National Museum of Ireland.
Lagore crannog. One-piece plain ' Iron Age
' collar from
among the old finds.
Courtesy National Museum of Ireland.
Plate V
Lagore crannog. Close-up of broken ends of the decoration of a
second composite collar. (See also Plate III, chain a.)
Courtesy National Museum of Ireland.
This content downloaded from 5.68.165.228 on Mon, 29 Jun 2015 16:40:56 UTCAll use subject to JSTOR Terms and Conditions