Poolewe: The last Bronze Age hoard in Scotland?

POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 1 Poolewe: The last Bronze Age hoard in Scotland? Matthew G. Knight, Dot Boughton and J. Peter Northover This is an Accepted Manuscript of an article published by Taylor & Francis in Archaeological Journal on 23rd November 2020, available online: https://www.tandfonline.com/doi/full/10.1080/00665983.2020.1824883. Fifty free e-prints are available to download here: https://www.tandfonline.com/eprint/PUQMEAYVKJAZXRSVCA8J/full?target=10. 1080/00665983.2020.1824883 POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 2 Poolewe: The last Bronze Age hoard in Scotland? Matthew G. Knighta*, Dot Boughtonb and J. Peter Northoverc a Curator of Prehistory, Scottish History and Archaeology, National Museums Scotland, Edinburgh, Scotland, United Kingdom EH1 1JF; b Archaeological small finds specialist, Independent; c Honorary Research Associate, School of Archaeology, University of Oxford * [email protected]; ORCiD: 0000-0001-6880-4417 In 1877, a hoard of nine copper alloy objects was recovered from a peat bog at Poolewe, Scotland, including axeheads, rings and an ornament. For the first time since its discovery, this article publishes the hoard in its entirety, including an assessment of typological features, full illustration and metallurgical analysis. Components of the hoard are characteristic of the British Llyn Fawr / Earliest Iron Age period (800–600 BC) suggesting the date of deposition, which is confirmed by a radiocarbon date from the wooden haft of an axehead. However, set in the broader context of Scotland and Britain during this period, it is suggested that this hoard in fact represents the last vestiges of the Late Bronze Age hoarding practice in Scotland. Keywords: hoard, Scotland, Bronze Age, Earliest Iron Age, Llyn Fawr, socketed axehead A hoard of nine bronze objects was excavated from a peat bog at Poolewe, Highland, Scotland during the late nineteenth century. It contained five socketed axeheads, three annular rings and a cup-ended ornament (Figure 1). Three of the axeheads fit typologically into the British Llyn Fawr / Earliest Iron Age period (c.800–600 BC), providing the earliest date of deposition, though aspects of the deposit suggest links with practices more common during the Late Bronze Age (c.1000-800 BC). It is the only definite hoard dating to this period from Scotland and has a long history in the National Museum of Scotland’s collections, yet it has received little dedicated research. In 2019 the hoard went on display at Gairloch Museum, making now an apt occasion to present POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 3 new or otherwise unpublished research on this group of objects. This paper details the history and contents of the hoard, its dating in light of new radiocarbon dates, and metallurgical analysis of the objects. We conclude by discussing the hoard’s position in Scotland and Britain during the Earliest Iron Age and highlighting that in fact it might represent the last traces of the Late Bronze Age hoarding tradition in Scotland. Figure 1: The Poolewe hoard. Photo: Neil McLean © National Museums Scotland History of discovery The hoard was recovered in 1877 from Poolewe, near Gairloch, by Mr Hector Maciver while digging peat. It was acquired by Sir Kenneth Mackenzie of Gairloch in 1879 and recorded in the Proceedings of the Society of Antiquaries of Scotland in 1880 by William POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 4 Jolly. The exact findspot is unknown; however, Jolly’s account provides a good indication: several Bronzes were come upon, on the high ground overlooking the River Ewe on the north side, some distance beyond the Public School… They were found 6 feet below the surface, all in one spot (Jolly 1880, 46). This helps confirm the authenticity of the group of objects as a hoard, and allows us to approximate the findspot. A peat bog south of the current Poolewe Primary School (i.e. the ‘Public School’ referred to by Jolly) provides the most likely findspot with the grid reference NG 863 804, now known as Lochan Dubh na Mòine (Figures 2 and 3). Lochan Dubh na Mòine forms a natural depression, which was originally a lochan before peat deposition, and still remains very wet and boggy, even during long, dry spells. Moreover, this spot has a significant depth of peat (at least 2m) and there is evidence of relatively recent peat cutting around the margins. The River Ewe flows on the north side of this findspot, and it is located relatively high in the area, making it a possible candidate for the reported ‘high ground’. The highest point is a nearby hill called Creag an Fhithich, though this has no evidence of historic peat cutting. Thus, we can be reasonably confident Lochan Dubh na Mòine is the approximate findspot.1 1 The authors are grateful to Jim Buchanan for his local knowledge and assistance in determining the probable findspot. POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 5 Figure 2: Probable findspot of the Poolewe hoard. Map (left): courtesy of Jim Buchanan. Figure 3: Oblique drone photograph of findspot of Poolewe hoard. Photo: Jim Buchanan. POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 6 In 1881, Sir Kenneth Mackenzie presented one of the nine objects, a cup-ended ornament, to the then National Museum of Antiquities of Scotland (Acc. No. X.DO 20). One of the annular rings was retained by the finder, Mr Maciver, and its whereabouts is now unknown. The location of the rest of the hoard after 1881 was unknown until 1958 when Professor John Coles rediscovered it in Conan House, Conon Bridge, the ancestral home of the Mackenzie family. The hoard was then loaned to the National Museum of Antiquities of Scotland, where it resided until 2019, when the hoard, including the cupended ornament, was redisplayed at Gairloch Museum. Contents The hoard comprised two complete and three incomplete socketed axeheads, three annular rings (one now lost) and a cup-ended ornament (Figure 4). All objects bear a similar bronze patina, which is the result of having been cleaned and stripped postrecovery; this has caused some surface degradation of all objects. Dimensions are recorded in millimetres. The typology for the axeheads follows Boughton (2015). 1. Socketed axehead, Type Sompting, Kingston variant Length 112.7; surviving cutting edge width 58.7; socket mouth dimensions (external) (width x height) 46.5x45.2; socket mouth dimensions (internal) (width x height) 34.1x35.5; weight 346.30g Largely complete cast socketed axehead with a sub-rectangular socket and a thick collar moulding. There are three vertical parallel ribs on both faces each terminating in a single pellet. The outer ribs descend from the moulding, though the central rib descends from a horizontal rib connecting the two outer ribs. The body of the axehead is narrow and expands to an incomplete broad crescentic cutting edge; no original edge survives and part of one face has fragmented. There is a small casting flaw near the thick side of the loop, but otherwise the axehead appears to have been cast well and prepared after casting. POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 7 Figure 4: The Poolewe hoard. Illustration Marion O’Neil © National Museums Scotland POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 8 2. Socketed axehead, Type Sompting, Tower Hill variant Length 112.2; cutting edge width 69.0; socket mouth dimensions (external) 44.0x41.9; socket mouth dimensions (internal) 29.5x31.0; weight 355.02g Complete cast socketed axehead with a circular socket and a thick bulbous collar moulding. The body is plain and undecorated and the sides gradually diverge from a narrow waist below the collar to a broad flared cutting edge. This edge is slightly dented, but it is difficult to relate this to evidence of use. Casting seams are prominent but show signs of some working and there is no casting material surviving around the socket. A piece of wooden haft (Quercus or Fraxinus sp., not heartwood) was preserved in the socket of this axehead measuring 66.4mm long (see Figure 1). This wood has evidence of charring, suggesting fire-hardening, and is wedge-shaped with a rectangular section to fit into the socket. 3. Socketed axehead, type uncertain Surviving length 32.3; cutting edge width 68.6; width at break 48.2; thickness at break 13.8; weight 92.12g Cutting-edge fragment of a cast socketed axehead, broken about 10mm above the internal aperture, probably in antiquity. There are no signs of casting flaws in the broken metal, though the core was at a slight angle during casting, creating thicker and thinner parts of the socket walls, the thinnest being 0.8mm and the thickest 4.5mm. There is a shallow circular depression (about 10mm diameter) on one face associated with the breakage, causing bowing of the metal in this area. The cutting edge is broad and slightly curved, indicating a Sompting type. The edge appears to have been worked and there are minor chips, which could indicate use-wear. 4. Socketed axehead, Type Sompting, Tower Hill variant? Length 92.5; cutting edge width 63.1; weight 167.42g. Socket dimensions are not observable. Incomplete body and cutting edge of a miscast socketed axehead. The surviving upper body indicates it would have possessed a sub-rectangular socketed mouth with a single bulbous collar moulding. The body of the axehead is plain and the sides diverge to a POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 9 slightly curved cutting edge. The axehead is incomplete as a result of failing to fill the mould during casting, though additional fragmentation has occurred since removal from the mould. The overall casting quality is poor and the surface is very porous. 5. Socketed axehead, uncertain or possibly Transitional Length 57.7; cutting edge width 37.2; weight 40.61g. Socket dimensions are not observable. Incomplete small cast socketed axehead, broken across half of the socket and down one face of the axehead. Evidence of a single bulbous collar moulding survives, and the remaining body is plain with a crescentic cutting edge. The surviving surface is pitted and porous and there is a black concretion residue on the inside of the socket which probably represents some of the original burnt core material. On the exterior surface around the side loop there is a slightly silvery patina which seems to represent the original uncleaned surface. 6. Annular ring, possibly harness fitting External diameter 47.2x48.3; internal diameter 30.3x30.3; width of ring 9.1; thickness of ring 8.3; weight 29.18g Complete circular annular ring with a hollow circular section and a sub-rectangular perforation. Small casting flaw in the exterior surface on one side. 7. Annular ring, vessel ring handle Class B2 (Gerloff 2010, No.61) External diameter 83.7; internal diameter 65.3; width of bar 6.6 / 3.6; thickness of bar 9.7; weight 53.12g Complete cast circular annular ring with an inverted T-shaped section and covered with a pitted black and bronze patina. 8. Cup-ended ornament / “dress-fastener” Maximum width of ornament (terminal to terminal) 118.6; bar width 6.1; bar thickness 4.4; terminal diameter (width x height) 51.4x54.3 and 51.8x52.6; weight 61.27g POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 10 Complete cast penannular bar ornament with concave circular terminals. The bar of the ornament has an oval cross-section and expands smoothly out to the terminals with no signs of a join on the exterior. The centre of the interior of the terminals has a raised lip. There is some dark grey concretion around the centre of each terminal which appears to represent the original patina. There are two fragmentary holes in each terminal where the metal is thinnest; this could be related to corrosion or, alternatively, excessive wear. 9. Annular ring, now lost Details of this object are unknown, but it was described as ‘similar to these [the surviving rings]’ (Jolly 1880, 46). Previous authors have assumed this was the same as the vessel ring handle (e.g. Gerloff 2010, 178; Boughton 2015, 239), but it could equally be similar to the hollow annular ring. Typologies and features of individual finds Socketed axeheads Of the five socketed axeheads in the Poolewe hoard, three, possibly four, are Somptingtype axeheads or variants of, and one is a small specimen of Transitional type. Sompting-type axeheads are the most characteristic Earliest Iron Age axehead type produced and used around 800–600 BC (Boughton 2015, 38–42). They are the largest of the entire corpus of Late Bronze Age and Earliest Iron Age socketed axeheads; whilst other Earliest Iron Age socketed axeheads are often small, light, thinly cast and have a high tin-content (11.45-22.96%) (Northover 1987, 186-187), the axeheads described as ‘Sompting’ are comparatively large and heavy (Boughton 2015, 38). One of us (DB) has recently reassessed the definition of Sompting axeheads presented by Colin Burgess (1971, 268) and key characteristics include their broad cutting edge (c. 6070mm), their overall shape and size, and their heavy weight (c.400g) (Boughton 2015, 150–152, figs 5.62–5.65; Burgess 1971, 268). The Poolewe examples all possess these features. Transitional axeheads are defined by slightly more elaborate decoration (Late Bronze Age socketed axeheads are mostly plain or ribbed) or a slightly bigger than POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 11 average size (e.g. 300-400g) (Boughton 2015, 94–101). In general, they share features of both Late Bronze Age and Earliest Iron Age axeheads. Although the Poolewe example is small and its socket damaged, it can be confidently placed within this type category due to its uncharacteristically splayed blade. All the axeheads showed signs of post-casting preparation, including the filing of casting sprues around the socket mouth and the casting seams down either side of the axeheads. This has even been performed on Axehead No.4, which is an incomplete miscast Sompting form. Despite having an incomplete socket and no side loop, the casting flash has been removed and it has a widely splayed, sharpened cutting edge. Postrecovery cleaning obscures definite signs of use-wear on the axeheads, but it is likely that some if not all were used prior to deposition. The surviving piece of hafting in Axehead No.2 illustrates this possibility, indicating that the axehead would have been a functional tool and may have been deposited with its haft. The complete nature of Axehead No.2 contrasts with the cutting-edge fragment of a socketed axehead (No.3), which was deliberately fragmented. Recent work on the deliberate destruction of socketed axeheads has demonstrated that breakage such as this was achieved by heating and striking the axehead with a blunt object (Knight 2017). Close analysis of the fragment revealed a faint circular hammer mark at the point of breakage on one face (Figure 5); furthermore, the socket wall is slightly bowed inwards at this point, all indicating intentional fragmentation. The deliberate fragmentation of axeheads, and indeed other bronze implements, was common in the Ewart Park phase of the Late Bronze Age (Knight 2017; Turner 2010) but is less common with Llyn Fawr period metalwork (Boughton 2015, 209–210, 264). The wooden haft from Axehead No.2 has been sampled twice for radiocarbon dating: once in the mid-1990s and again in 2018. Both samples were processed by SUERC for radiocarbon dating by AMS, with the graphite target for the first sample measured at the University of Arizona (AA-). AA-29710 (2465 ±40 BP) was processed following pre-treatment methods detailed in Stenhouse and Baxter (1983) and graphitized following the method described in Slota et al. (1987). The sample was measured by AMS following Donahue et al. (1997). Using the internationally-agreed calibration curve (IntCal13) of Reimer et al (2013), the maximum intercept method (Stuiver and Reimer POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 12 Figure 5: The hammer blow on the fragmented socketed axehead. Photo: M. Knight © National Museums Scotland 1986), and the OxCal computer program (Bronk Ramsey 2009), the result is calibrated to 780–400 cal BC (95% confidence). The second sample (SUERC-81222) was processed in 2018, following the methods of Dunbar et al. (2016), and it provided a radiocarbon age of 2553 ±21 BP (Knight 2019a), which calibrates to 800–560 cal BC (95% confidence), with the 68% confidence interval at 790–760 cal BC (Figure 6). The two results are statistically consistent (T’=3.8; df=1; T’(5%)=3.8; Ward and Wilson 1978) and consistent with the expected typology of the axehead, with a high likelihood that it was hafted, and perhaps deposited, in the first half of the 8th century BC (D. Hamilton pers. comm.). POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 13 Figure 6: The calibrated date on the wooden axe haft fragment (SUERC-81222) (provided by SUERC, University of Glasgow) Perforated annular ring The Poolewe annular ring is an unusual form but is consistent with the production of solid and hollow annular rings in the Late Bronze Age and Earliest Iron Age. The annular ring shows no signs of wear, though we must accept some of this may have been removed through cleaning. Upon initial inspection, it was posited that this object may have functioned as a terret or piece of horse gear, but the lack of observable wear on the interior of the ring, even accepting the effects of cleaning, does not indicate reins having passed through it. It could have alternatively functioned as a mount or accoutrement. Vessel ring handle(s) The vessel ring handle typologically fits within Gerloff’s Class B2 series, broadly dating to c.675–600 BC (2010, 181–183). Although it cannot be conclusively attributed to either a bucket or a cauldron, Gerloff notes its similarity in size to the Class B2 cauldron ring POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 14 handles in the Sompting hoard, Sussex, as well as highlighting the current absence of bucket handles with similar cross-sections (2010, 178). As with the annular ring, the Poolewe vessel ring handle shows limited signs of wear. Cup-ended ornament Cup-ended ornaments (alternatively ‘dress fasteners’) are almost exclusively produced in gold during the Late Bronze Age in Britain and Ireland (c.1000–800 BC) (Eogan 1994, 88). Whilst eighteen gold examples are known from Late Bronze Age hoards and as single finds in Scotland (Eogan 1994, 142–144), the Poolewe ornament is the only confirmed copper alloy version known (see below). It is therefore anomalous for its production in bronze and its presence in an otherwise Earliest Iron Age hoard. The surface of the Poolewe ornament appears to have been damaged by post-recovery cleaning, particularly where it is thinnest at the terminal, though this may also represent a degree of wear raising the possibility that it may have been an already old object when deposited. This point is explored further below. An important aspect of the bronze cup-ended ornament is its manufacture. Maryon suggested that gold cup-ended ornaments in their simplest forms were cast, but alternative examples with larger cup terminals, such as those in Ireland, required the terminals to be produced first and then the bar of the ornament was cast onto the terminals (1938, 201-203). Alternatively, the components may be produced separately and attached by soldering or ‘burning together’ the different elements, creating a seamless join (Maryon 1938, 201-203). Conceivably, such a method could be used for producing a copper alloy counterpart. The bar of the Poolewe ornament seamlessly expands into the terminals and there is no evidence that the terminals were hammered out (Figure 7), so it is possible that these were cast onto the bar using a lost-wax casting process. Further analysis of this object involving microscopy or x-ray was not possible as part of the present research but would help clarify aspects of manufacture in the future. POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 15 Figure 7: The bronze cup-ended ornament. Photo: Neil McLean © National Museums Scotland Metallurgical analysis by J.P. Northover The metallurgical study of the Poolewe hoard took place in two stages. The first consisted of the analysis of the vessel ring handle (No. 7) as part of a programme of metallurgy for inclusion in a catalogue of Atlantic cauldrons and buckets (Northover 2010), while the second comprised the analysis of six of the other seven locatable objects. This was carried out as part of a project of characterising Scottish Late Bronze Age hoards, consequent upon the detailed study of the St Andrew’s, Fife, hoard (Cowie, Northover and O’Connor 1998). Objects were selected on the basis of a recognisable typology meaning that the axehead fragment (No. 3) was excluded. All samples were drilled, using a 1mm diameter drill bit, hot-mounted in a carbon filled thermosetting resin, and ground and polished to a 1µm diamond finish. Analysis was by electron probe microanalysis, for twelve elements in the case of the ring handle (No. 7), and thirteen, including sulphur, for the other objects; detection limits were 100200ppm. Three analyses were made on each sample; the mean compositions, normalised to 100%, are presented in Table 1; all concentrations are in weight %. POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? Table 1: Compositional analysis of the Poolewe objects Hoard Sample Object Fe Co No. 1 Ed620 Socketed axehead, 0.00 0.02 Sompting 2 Ed640 Socketed axehead, 0.07 0.05 Sompting 4 Ed611 Socketed axehead, 0.02 0.00 fragmentary 5 Ed618 Socketed axehead, 0.06 0.00 ?Transitional 6 Ed614 Hollow annular 0.06 0.06 ring 7 SIMG8 Vessel ring handle 0.04 0.03 9 8 Ed726 Cup-ended 0.01 0.01 ornament 16 Ni Cu Zn As Sb Sn Ag Bi Pb Au S 0.07 86.3 3 88.8 3 94.0 6 88.6 5 87.5 4 87.3 9 89.8 2 0.00 0.52 0.02 0.03 0.25 2.50 0.01 0.02 0.00 0.20 0.12 0.06 0.00 0.32 0.01 0.02 0.00 0.29 0.11 10.2 2 10.2 0 4.80 0.05 0.02 0.49 0.00 0.01 0.01 0.27 0.17 9.48 0.12 0.01 1.08 0.00 0.03 0.00 0.09 0.07 0.03 0.01 1.67 0.00 0.10 0.00 0.20 0.00 0.03 0.01 0.60 0.00 0.01 0.16 0.05 10.2 6 12.2 7 9.33 0.06 0.03 0.39 0.07 0.14 0.14 0.12 0.10 0.03 0.03 0.02 POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 17 All the analysed objects were cast in medium-tin, unleaded or low-lead bronze, except for the miscast socketed axehead (No. 4) which is in a low tin bronze with 4.8% tin and 0.5% lead. The seven analysed objects can be grouped into five impurity patterns, two pairs (Nos 4 and 5, and Nos 2 and 6) and three individual pieces (Nos 1, 7 and 8), a variety that shows at once that the metal supply had not become homogenised by repeated recycling and that new metal was entering the system. The first pair (Nos 4 and 5) has an As/Sb/Ni/Ag impurity pattern, but the tin contents differ widely at 4.8% and 9.5%. This pattern indicates that at least some portion of the copper in the alloy is of Continental origin (Northover 1983). The same is true of the second pair (Nos 2 and 6), but here there is a cobalt impurity as well (0.05-0.06%). This is something very unusual in Britain but can be found elsewhere. The comprehensive programme of analysis carried out for the lakeside settlement of ZugSumpf in Switzerland (Northover 2004) showed a consistent presence of As/Sb/Co/Ni/Ag impurity patterns with a wide range of impurity concentrations, with Group 9d in the Zug-Sumpf corpus matching these very well. One of the complete Sompting axeheads (No. 1) has arsenic and bismuth as its principal impurities, atypical for the Late Bronze Age but very typical of Anglesey and Cornish copper at the end of the 18th century BC, so a British origin for the copper in this axehead might be proposed (Northover and Wilcox 2013). There is some similarity between the compositions of the vessel ring handle (No. 7) and the cup-ended ornament (No. 8) and some connections might be suggested by the 0.03% each of cobalt and nickel in the ring handle. A very similar pattern is seen in the base sheet of the cauldron from Kincardine Moss (Gerloff 2010, No. 69), but with 0.05% cobalt and 0.06% nickel, while the impurities in the top sheets are close to those of the cup-ended ornament. The Kincardine Moss cauldron, one of Gerloff’s Class B3 cauldrons, Type Kincardine Moss, is seen as representing some hybridisation between Atlantic cauldrons of Class B2 from the Hallstatt C (Llyn Fawr) period and Continental cauldrons of Ha D (Gerloff 2010, 186–189). There is, though, no evidence for what handle arrangements B3 cauldrons had, if any, so it is not possible to associate the Poolewe ring handle with such a cauldron. To place the results from Poolewe in a broader context two sets of comparisons were made: the first with other hoards from the mainland of Britain from the Llyn Fawr POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 18 period (Figure 8), and the second with all Late Bronze Age hoards from Scotland (Figure 9). The simplest means of visualising these is through bivariate scatter plots; this is especially the case with the impurities, where concentrations are relatively low and some element pairs are strongly correlated, thus limiting the usefulness of multivariate statistics. Because the number of objects from Poolewe is small compared with the total number of analyses in each plot, the symbols for Poolewe have been enlarged to aid clarity. For the Llyn Fawr period the comparisons are only between axeheads and axehead fragments since they occur in all the hoards and are in the majority in most.2 The comparisons are made graphically, with plots of lead against tin in Figure 8a and nickel against antimony in Figure 8b. The disadvantage in this comparison is that the other hoards are in south-central and south-west England and south-east Wales. Regarding the alloy used there is a clear difference between the Tower Hill, Oxfordshire, and Figheldean Down, Wiltshire, axeheads in terms of tin content and between King’s Weston Hill, Bristol, and Figheldean Down in lead content, even though the findspots are relatively close to each other. Poolewe matches Leckwith Moors, Vale of Glamorgan, and King’s Weston Hill most closely, except for the outlier with a low tin content, which is closer to Tower Hill. To discover if a similar patterning could be observed through the impurity patterns, a plot was made of nickel against antimony (Figure 8b). With copper from Continental Fahlerz sources nickel and antimony can be strongly correlated so this plot can be used to differentiate between this and other types of source. The result is that we see again a separation between the Tower Hill and Figheldean Down hoards. As might be expected from different impurity groups the Poolewe data are split between three axeheads which have similarities to Tower Hill axeheads, but with higher nickel and antimony contents, and a single axehead which is close to Figheldean Down and King’s Weston Hill axeheads, and also some similarity to some axeheads in the hoard from Mylor, Cornwall. It is important to note, though, that the alloy contents and impurity patterns do not correlate for the Poolewe data with the low tin bronze being associated with the copper with the highest impurity content. 2 Much of the data was compiled for the publication of the Tower Hill, Oxfordshire hoard (Coombs, Northover and Maskal 2004); the rest come from the writer’s own database. POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 19 Figure 8: Comparisons of trace elements of socketed axeheads from select Earliest Iron Age / Llyn Fawr copper alloy hoards in Britain with Poolewe: (a) lead and tin; (b) nickel and antimony. POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? Figure 9: Comparisons of trace elements from objects in Late Bronze Age hoards in Scotland with Poolewe: (a) lead and tin; (b) nickel and antimony; (c) antimony and arsenic. 20 POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 21 The second question posed was the relationship of Poolewe to the overall sequence of Scottish Late Bronze Age metalwork (c.1150-800 BC). In Figures 9a-c the seven Poolewe analyses are plotted against the totality of analysed Scottish Late Bronze Age bronze for lead and tin, nickel and antimony, and antimony and arsenic. In Figure 9a five of the alloys fall centrally in the Scottish distribution, with the vessel ring handle having a slightly higher tin content; the low tin bronze axehead is an outlier for any stage of the Late Bronze Age. Turning to the impurities, nickel and antimony and antimony and arsenic can be strongly correlated (Figures 9b and 9c). The Poolewe data fall off these trends, all the antimony concentrations being below both. There has clearly been a change in the metal supply circulating in Scotland. This change has not been explored to any great extent in Britain, but there is another area of northern Europe where it has: Denmark (Liversage 2000). The corresponding change in Scandinavia is from Period V to Period VI and here, also, there is a modest trend towards lower antimony contents in Denmark. For Scotland, Poolewe may reflect a similar change. This is not to imply any direct connection between Denmark and Scotland, rather that developments in the ultimate source area for much of the copper influences the outcome in both Scotland and Denmark in a similar way. The implications of this change in the metal supply are discussed below. The objects in context Having established key features of the individual objects and their metallurgy, they can be contextualised alongside similar objects from Britain to establish their individual significance. The hoard is then compared with Late Bronze Age and Earliest Iron Age hoards and set within the wider context of Scotland and Britain. Socketed axeheads Considering the entire corpus of Earliest Iron Age metalwork from Britain dating from c. 800–600 BC, the predominant artefact type is without doubt the socketed axehead. More than 1400 Earliest Iron Age socketed axeheads and at least six moulds are now known from Britain (Boughton 2015, 2). Forty-seven socketed axeheads are from Scotland, ten from three possible associations (including the five from Poolewe) and 37 as single finds; POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 22 at least four moulds are known from three sites in Scotland (Figure 10). The nearest broadly contemporary deposit to Poolewe is an iron socketed axehead, now lost, reported from Inveran, south of Poolewe on the edge of Loch Maree (MacKie 1967), probably dating to the Llyn Fawr period, or perhaps slightly later. However, compared to the region’s actual size, Scotland has only a very small corpus of Earliest Iron Age socketed axeheads (Boughton 2015, 319–321), making Poolewe a significant concentration of the overall. The number of single finds of Earliest Iron Age socketed axeheads (37) is comparable to that of Boughton’s Southern region (Hampshire, Dorset, Wiltshire, Isle of Wight) (33) but the density of finds is much higher in the south than it is in Scotland (Boughton 2015, 242f., 322). The unassociated socketed axeheads show a wide distribution across Scotland, but there are very few notable clusters and concentrations of specific axehead types. The associated axeheads and single finds share one common trait, however, which is that their individual findspots are either in the Scottish Lowlands or quite close to the coast. There are few finds from the Scottish Highlands and even in southern Scotland none of the axeheads were found on higher ground or in the upland areas. The location of Poolewe so far north is therefore unusual. Its location on the west coast, however, is perhaps less anomalous when we consider the small cluster of single axeheads further south in Argyll and Bute. Two ceramic mould fragments for producing Sompting axeheads recovered from Little Dunagoil, Bute (Coles 1960, 89), suggest this may also have been an area in which these axeheads were produced. Alongside this we can consider two sets of complete bivalve stone moulds for Sompting axeheads from Rosskeen in the eastern Highland region (Wilson 1863, 345–346).3 Coles (1960, 36–37) suggests that in the light of the discovery of three sets of moulds for casting decorated Sompting axeheads, Scotland was probably not only a recipient of such axeheads, but also a producer. This is supported by recent finds of moulds for Earliest Iron Age socketed axeheads from the prehistoric settlement at Bellfield Farm, North Kessock, Highland, on the east coast alongside a range of other 3 Due to a complex history post-recovery in which they were separated, these two mould sets have previously been erroneously disassociated and attributed very different provenances, one from Ardrossan, Ayrshire, and one from Stittenham, Rosskeen. It is clear from Wilson’s (1863, 345–346) account and illustrations that these two moulds are in fact from the same site in Rosskeen. POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 23 Figure 10: A map of Earliest Iron Age / Llyn Fawr axeheads and moulds from Scotland. Sites mentioned in text are labelled. Finds with an imprecise findspot (e.g. ‘Scotland’) have not been mapped. POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 24 object moulds (Clark et al. 2017; see below). In connection with this, it is noteworthy that only two moulds for Sompting-type axeheads are known from England: Melcombe Horsey, Dorset, and Ham Hill, Somerset (Boughton 2015, Nos 640 and 1409). However, no moulds have been found in regions with the greatest concentrations of Sompting-type axeheads: the Thames Valley and East Anglia (Boughton 2015). In northern England, there is metalworking evidence from the Furness peninsula, Cumbria, where a piece of casting residue was found in a hoard with three socketed axeheads at Ulverston, Cumbria (Boughton 2015, Nos 1395-1397). Furthermore, the two decorated axeheads from Ulverston were made in the same mould or from the same mould template as one of the axeheads from an older hoard from the Furness (Skelmore Heads, Cumbria) and a single find from Dunnichen, Angus (Boughton 2015, 320); even though no mould was found in the Furness, the metalworking evidence suggests that all four axeheads were made in the Furness peninsula. This allows us an insight into the different ‘histories’ axeheads may have. One of the axeheads from Ulverston and the axehead from Skelmore Heads were deposited in ascast conditions while the axehead from Angus shows clear signs of wear and resharpening on both the upper and lower surfaces of the body (Boughton 2015, Nos 193, 1245 and 1395). This suggests that it probably had a longer ‘life’ than the two axeheads from the Furness which were never finished. One can draw parallels here with the miscast axehead in the Poolewe hoard showing signs of use-wear. Similarly, at Shelford, Nottinghamshire, a miscast Late Bronze Age–Earliest Iron Age axehead had been prepared for use despite its imperfect form (Boughton 2015, No. 930). This contrasts with other complete axeheads from Earliest Iron Age contexts that have not been prepared or used. The contexts in which Earliest Iron Age socketed axeheads were deposited, which includes places near settlements, coasts and hilltops, infer no overall preference (Boughton 2015), but across Britain, places in or close to watery locations are common contexts, a trend that continues from the Late Bronze Age (Boughton 2015, 321; Bradley 1998 [1990]; O’Connor 2007a). In Scotland, most depositions of Earliest Iron Age axeheads were made near the sea or a river or loch, though, apart from Poolewe, their POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 25 recovery from within wetland locations is unknown;4 thus, while the deposition of the Poolewe hoard in proximity to a sea loch broadly conforms with the expected practice for depositing axeheads between 800–600 BC, its deposition in a peat bog, possibly in or near a former lochan, stands out in Scotland. By contrast, Late Bronze Age hoards and depositions in Scotland have commonly been recovered from peat bogs, often including axeheads, such as an axehead and gouge from Achnahannait, Skye, or the hoard of axeheads, tools and ornaments from Adabrock, Harris (Coles 1960, 110, 127). Moreover, Jolly’s (1880) original account of the Poolewe hoard describes earlier Bronze Age metalwork found nearby, including axeheads and spearheads. This suggests the Poolewe hoard was deposited according to widely understood practices and in a landscape that had a long history of deposition; practices in this area may have been undertaken according to past traditions, thus reflecting the continuation of important concepts from the period 1100-800 BC and including new forms of objects, such as Sompting axeheads. Well-dated contexts for Sompting-type axeheads are rare, particularly in northern Britain, and dating of these axeheads has mainly been performed using typological analysis extrapolated from the few well-associated hoards and by comparisons with artefacts from Britain and continental Europe. This emphasises the significance of the Poolewe dates, as these represent the only direct radiocarbon dates from an object within a Llyn Fawr hoard in Britain. The only other radiocarbon date directly associated with a Sompting axehead comes from the River Thames (see Table 2), though dated contexts from which Sompting-type axeheads have been recovered infer their continued production and use throughout the Llyn Fawr period. A fragment of a possible Sompting axehead fragment was recovered from the basal accumulation layers of the midden at East Chisenbury, Wiltshire, which produced radiocarbon dates suggesting deposition during the ninth–eighth centuries cal BC (Waddington et al. 2018, 29). At the Llanmaes midden site, Vale of Glamorgan, fragments of Earliest Iron Age/Early Iron Age metalwork, including multiple Sompting and Armorican axeheads, were probably deposited c.675–530/520 BC based on typo-chronological associations and other radiocarbon dates from the site (Gwilt et al. 2016, 309–315; A. Gwilt pers. comm.). Other One exception may be a Sompting-type axehead from ‘Corsbie Tower’, which may have been recovered from a nearby peat bog (Knight 2019b, 35). 4 POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 26 hoards and associations, such as the eponymous Sompting hoard, similarly suggests that Sompting axeheads were in use into the late seventh century BC (Boughton 2015; O’Connor 2007a, 71–73). The radiocarbon date associated with the Poolewe example thus infers that its hafting and use falls relatively early in the overall development and circulation of Sompting axeheads. Table 2: Radiocarbon dates directly associated with Earliest Iron Age / Llyn Fawr metalwork in Britain Findspot Poolewe, Highland Broom, Warwickshire Buscott Lock, Oxfordshire Llanmaes, Vale of Glamorgan River Thames, Kew Deer Park, Middlesex East Chisenbury, Wiltshire Material dated Axe haft (Quercus or Fraxinus sp., not heartwood) Charcoal (Fraxinus sp., heartwood) Charcoal (Fraxinus sp., heartwood) Axe haft Pig metatarsal Axe haft (Salix/ Populus) Animal bone (sheep/goat radius) Carbonised residue from pot sherd Associated metalwork Context BP Hoard 2553±21 Sompting axehead Class B2 cauldron fragments Pyre site with remains of cremation deposits and charcoal Iron axehead Class B2 cauldron ring handle fragment Sompting axehead Single find from river Posthole (F076) of pre-midden roundhouse Single find from river Possible Sompting axehead fragment Midden basal accumulation layer (Trench A, context 11) cal BC (2σ, 95.4% probability) 801–571 2465 ±40 780-400 2570±55 833–520 Lab code Reference SUERC81222 AA29710 Knight 2019a; this paper OxA6282 Gale in Palmer 1999, 53 2475±55 773–416 OxA6283 2480±50 776–416 2492±33 782–418 OxA6216 UB– 6428 Needham et al. 1997, 98 Gwilt et al. 2016, 301, 310, 322 2545±55 830–510 OxA4658 Needham et al. 1997, 72 2682±29 897-802* OxA20277 2521±28 794–739 (30.3%) 692–659 (14.9%) 651–543 (50.2%)* OxA20217 Waddington et al. 2018, 23 *Dates calibrated by authors using the OxCal computer programme, v.4.3 (Bronk Ramsey 2009), calibration curve IntCal 13 (Reimer et al. 2013) for this paper. POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 27 Perforated annular ring The function and significance of the perforated hollow ring is difficult to determine, especially as few parallels exist. Whilst solid and hollow annular rings are known from Late Bronze Age and Earliest Iron Age hoards, few are perforated. Hawkes highlighted similarities between the Poolewe ring and the perforated hollow rings in the Late Bronze Age hoard from Trillick, Co. Tyrone (Hawkes and Clarke 1963, 227). This hoard contained eight annular rings, a sunflower headed pin, a socketed hammer and a socketed axehead (Eogan 1983, 161–162). Two of the rings are perforated with two vertical holes opposite each other through the body from the outer edge to the inner edge, the function of which is unclear. A similarly perforated ring was found in the Derryhale hoard, Co. Armagh, associated with another bronze cup-ended ornament (Eogan 1983, 57–60; see below). The alternative perforation arrangement on the Poolewe ring (i.e. horizontal rather than vertical) means we cannot assume that these rings all served the same general function though. A complex ring with eight horizontal perforations and a D-shaped suspension loop was recovered as part of the Earliest Iron Age hoard from Tower Hill, Oxfordshire, found with socketed axeheads, ornaments and copper alloy fragments, though the original function of this ring is unclear (Coombs et al. 2004, 206, 215, figure 11.14) and it significantly differs from the Poolewe ring. Although dating much later, a Roman harness fitting from Horncastle, Lincolnshire, is remarkably similar in form (Staves 2006), suggesting that the Poolewe object could indeed be considered a harness fitting, perhaps for a horse.5 Horse gear, including harnesses and cart fittings, are generally very rare in British Late Bronze Age and Earliest Iron Age contexts; notable exceptions include multiple hoards from Wales (Parc-y-Meirch, Denbighshire; Leckwith Moors, Cardiff; and the eponymous Llyn Fawr assemblage, Vale of Glamorgan [O’Connor 1980, 419–420, Nos 217, 218; Savory 1980, 119, 187; Thomas 1989, 281]) and the collection of Late Bronze Age terret rings and fittings from Horsehope, Scottish Borders (Piggott 1953). None of this material matches the ring from Poolewe though and, importantly, there are no signs of wear on the Poolewe ring that one might expect with a terret or harness. Nonetheless, the ring probably functioned as an accoutrement of some kind. 5 The authors are grateful to Anna Lewis for highlighting this comparison. POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 28 Vessel ring handle(s) Cauldrons and buckets are viewed often as feasting vessels, though many of the Class B cauldrons (from which the Poolewe handle may derive, specifically Class B2) may have never been used, serving only as symbolic objects (Gerloff 2010, 196). Class B cauldrons were massive and difficult to move, but if used for cooking and feasting they would have fed a large number of people; the inclusion of a vessel ring handle in the Poolewe group may therefore infer local communal cooking activities. Evidence for such feasting events are lacking in Scotland but are most clearly seen at midden sites in southern Britain. Excavations at the Llanmaes midden and pre-midden site revealed fragments of multiple cauldrons, cups and ladles in association with swan’s neck pins and socketed axeheads, deposited amongst animal remains and other bronze debris (Boughton 2015, 194; Gwilt et al. 2016). The midden accumulated over time and may represent a site of communal gathering or feasting, similar to the middens at Broom, Warwickshire, and East Chisenbury and Potterne, both Wiltshire (Lawson and Allen 2000; McOmish, Field and Brown 2010; Watson 1999). At Llanmaes, multiple fragments of Class B2 vessels were found within the midden, as well as an incomplete vessel ring handle deposited in the posthole of a pre-midden roundhouse associated with a pig metatarsal, which produced a radiocarbon date of 782–418 cal BC (see Table 2; Gwilt et al. 2016, 310). At Broom, Class B2 cauldron fragments were also found with associated dates in the Llyn Fawr period or perhaps slightly later (see Table 2; Gale in Palmer 1999, 53). Gwilt et al. (2016, 309-310) suggest a depositional date for the Class B2 vessel fragments at Llanmaes in the seventh century BC (675-600 BC) based on typo-chronological grounds and radiocarbon dates from the lower midden layers; the vessel ring with associated radiocarbon dates indicates use of vessels prior to the midden formation (see also Gerloff 2010, 181-183). However, the associated date from Poolewe suggests an earlier production date for Class B2 vessels in the eighth century BC. This would overlap with typologically earlier Class B1 cauldrons, as suggested by Gerloff (2010, 182). As with the Sompting axeheads, it is probable Class B2 vessels had a long sequence of production and circulation throughout the period 800-600 BC, with the Poolewe example occurring towards the beginning of this sequence. POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 29 In Scotland, relatively few vessels are known from the Late Bronze Age and Earliest Iron Age, with the earliest known example deposited at Hattonknowe, Scottish Borders, c.1100–1000 BC (Buchan 1905). A largely complete bucket from Flanders Moss, Stirling, could date to c.800 BC (Gerloff 2010, No.69). More commonly, vessels are indicated by ring handles, such as those from the Duddingston Loch assemblage, Edinburgh, and the hoards from Dowalton Loch, Dumfries and Galloway, and Dalduff, South Ayrshire (Gerloff 2010, 90f., 272–274, Nos 26, 126 and 127), all dating to c.1000800 BC. There is none definitely known from the Llyn Fawr phase in Scotland. We can thus assume that the use of large vessels, perhaps for communal feasting events, was well-established, albeit uncommon, in Scotland by the time the Poolewe hoard was deposited. However, the single vessel ring from Poolewe raises an obvious question: where is the rest of the vessel? The deposition of fragments or incomplete objects is common in the Late Bronze Age, as highlighted by the Poolewe axehead cutting-edge fragment. Moreover, the deliberate removal or fragmentation of ring handles from vessels has been observed on multiple Late Bronze Age–Early Iron Age cauldrons across Britain and Ireland, as at Llanmaes (Gerloff 2010, 196; Gwilt et al. 2016). One of the handles on the Hattonknowe cauldron, for instance, had been deliberately removed, and the ring handles in the Scottish hoards mentioned above may also have been intentionally separated from the vessels. Depositing pairs of vessel rings and staples is occasionally seen across Britain, such as in the Dalduff hoard, strengthening the possibility that the missing annular ring from Poolewe was indeed another vessel ring. Reasons for deliberately removing and depositing pairs of vessel rings are difficult to interpret. The deposition of partial objects may represent a pars pro toto practice, in which part of an object is deposited as a token of the original complete object (cf. Hansen 1994; 2016; Fontijn 2008; Knight 2018, 35, 433f.). However, such a theory does not necessarily explain the combination of complete and incomplete objects in the hoard, nor the variety. Cup-ended ornament Eighteen gold cup-ended ornaments are known from Scotland (Eogan 1994, 142–144), frequently deposited with other gold objects, as at Heights of Brae, Highland (Clarke and POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 30 Kemp 1984), or else as single finds. By contrast, the Poolewe cup-ended ornament is the only copper alloy example from Scotland and is one of only five possible examples known from Britain and Ireland (Table 3). Another possible bronze example is recorded to have come from an unspecified location attributed to ‘Scotland’ (Coles 1960, 91; Wilson 1863, 461), though no further information is known and, although Wilson notes it was donated to the (then) National Museum of Antiquities of Scotland, it can no longer be located. Furthermore, Wilson’s account of this object compares it to a small ‘dilated penannular ring’ (1863, 461) found with a gold cup-ended ornament from Galway, Ireland. The smaller Galway ornament is better classed as a sleeve-fastener (Eogan 1983, 90-91); overall it seems unlikely the object in question is a true bronze cup-ended ornament. A miniature bronze ornament recovered through metal-detecting from Powys seems to represent a small variation of a bronze cup-ended ornament (Lodwick 2010) and mimics the three-piece manufacture of larger gold examples where the terminals are made separately (M. Cahill pers.comm.). The only secure association of another fullsized bronze cup-ended ornament is in the hoard from Derryhale, Northern Ireland (Eogan 1983, 57–60). This hoard contained an axehead, knives, annular rings, sunflowerheaded pins and glass and faience beads and can be relatively dated to the Irish Dowris phase (c.900–800 BC) equivalent to the British Ewart Park phase. This suggests that bronze cup-ended ornaments were likely in circulation at the same time as their gold counterparts, which are more numerous and date to c.1000–800 BC. This dating is important as it suggests that the Poolewe ornament may have been produced earlier than other objects in the hoard, but continued in circulation into the first half of the eighth century BC when the hoard was most likely deposited. If we accept that the Poolewe ornament was contemporary with gold versions, this raises two possibilities: the Poolewe ornament was produced late in its expected sequence (c.800 BC); or was of some age when deposited. Even if the former is the case, the radiocarbon date associated with the hoard implies that the ornament could have been several years, if not decades, old when deposited; this interpretation is strengthened by the possible wear on the terminals. This, and its typologically anomalous nature, suggests we are dealing with a group of objects that may have been accumulated over some time, a phenomenon that is increasingly recognised in the British Bronze and Iron Ages (e.g. Boughton 2019; Hingley 2009; Joy POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 31 2016; Knight 2019b; forthcoming). One hesitates to attribute an ‘heirloom’ status to the ornament, but it probably accrued some social value over time. Table 3: Certain or possible copper alloy cup-ended ornaments from Britain and Ireland Site Context Associated Notes Reference objects 1 Poolewe Hoard Socketed This paper axeheads, annular rings 2 Derryhale, Hoard Socketed Eogan 1983, 57–60 Co. Armagh, axehead, knives, Northern annular rings, Ireland sunflower pins, glass and faience beads 3 ‘Ireland’ Uncertain Eogan 1994, 144 4 Brecon Gaer, Single Miniature? Lodwick 2010 Powys, find [incorrectly listed as Wales Dymock, Gloucestershire] 5 ‘Scotland’ – Uncertain Probably not Wilson 1863, 461; provenance a true cupColes 1960, 91 uncertain ended ornament The Place of the Poolewe hoard in Earliest Iron Age Scotland and Britain The Late Bronze Age–Earliest Iron Age transition c.800 BC has often been viewed through the lens of the metalwork deposition (see Needham 2007). Large quantities of bronze objects were deposited as single finds and in hoards across Britain and Europe during the period 1150–800 BC. In Scotland, John Coles (1960) listed over 60 Late Bronze Age hoards, the majority dating to the early first millennium BC; to this we can add new discoveries in the sixty years since. A discovery of note is the Baile-a-Chladaidh hoard from Dores, near Inverness, on the edge of Loch Ness. This hoard contained two complete and two fragmentary socketed axeheads, a socketed gouge, a tanged chisel, one complete and one fragmentary socketed knife and two pieces of casting jet (Clark et al. 2017, 50). Although typologically earlier, its Highland location and the combination of complete and incomplete axeheads compares well with aspects of the Poolewe hoard. POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 32 The overall character of Late Bronze Age hoards in Scotland is similar, encompassing complete and/or fragmented weapons, tools and ornaments, largely composed of bronze, but also including gold, amber, ceramic and stone objects. The Balmashanner hoard, Angus, for instance, included a deliberately broken socketed axehead fragment, bronze armlets and rings, fragments of a cast bronze bowl, gold penannular rings, amber and cannel coal or shale beads, and an iron ring – the earliest known iron object in Scotland – buried in a ceramic vessel (Anderson 1892; Coles 1960, 98–99; A. Sheridan pers. comm.). The Adabrock hoard, Lewis, was similarly diverse, comprising bronze socketed axeheads, spearheads, tools and razors, as well as fragments of a decorated bronze vessel with continental affinities, whetstones, and beads made of glass, amber and gold (Coles 1960, 127). As at Poolewe, the Adabrock hoard was deposited in a bog, probably during the ninth century BC. Late Bronze Age hoards in the rest of Britain frequently contain heavily used, broken-up weapons, tools and ornaments (Taylor 1993; Huth 1997; Turner 2010). By contrast, the composition of Earliest Iron Age metalwork hoards is usually very different, often including complete artefacts in good or as-cast condition, mostly comprising axeheads, as well as items relating to communal feasting and horse gear and vehicle fittings (O’Connor 2007a; Boughton 2015, 176–177). Earliest Iron Age metalwork hoards have largely been recovered from southern Britain, with the biggest concentrations in the Wessex region of southern England and south-east Wales. The axedominant nature of Poolewe, associated with ornaments and a vessel ring handle can be compared with these broader compositional traits, despite being geographically very separate. Geographically the closest definite hoards are from Cumbria: Skelmore and Ulverston, comprising six and three complete socketed axeheads respectively, though no additional artefacts (Boughton 2015, Nos 193–198, 1395–1397). The hoard from Ferring, Sussex, deposited during the first half of the eighth century BC, contained one broken and nine complete socketed axeheads of various forms as well as fragments of a two Gündlingen swords, a socketed leather-working knife and part of a cast copper alloy belt or strap attachment (Boughton 2015, Nos 999–1008). The eponymous Sompting hoard, Sussex, contained two fragmentary cauldrons and 17 socketed axeheads (Curwen 1948); this hoard was deposited c.650-600 BC indicating a period of around 200 years during POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 33 which hoards of this nature persisted (O’Connor 2007a), continuing trends that began in the Late Bronze Age. However, unlike southern Britain, relatively little metalwork entered the ground in Scotland between 800–600 BC, except as single finds, most of which is concentrated in the south and east of Scotland, and included swords, razors and axeheads (O’Connor 2007a; Boughton 2015; see above). In addition, Poolewe remains the only definite hoard from Scotland dating to this period. Although the Balmashanner hoard contains an iron object, the treatment and diverse character of the associated objects, including bronze, gold, amber and cannel coal or shale ornaments, suggests a depositional date c.900–800 BC during the Late Bronze Age; the iron ring in this case represents an early inclusion of an iron object, rather than reflecting a Llyn Fawr-phase hoard. Two supposed associations date to the Llyn Fawr phase, from Lamancha, Scottish Borders, and Tillicoultry, Clackmannanshire (see Figure 10), though neither are convincing. Nothing is known of the find circumstances of three Armorican axeheads from Lamancha (Coles 1960, 124), and these lie outside the typical distribution of this axehead type in the furthest south and south west of England. This so-called hoard must be considered suspicious, and it is probable that the three Armorican axeheads were single antiquarian finds, having come either from the area or, more likely, from further south. Similarly, two decorated Sompting axeheads found near an earlier stone circle at Tillicoultry may have been found close together, but not directly associated (O’Connor 2007b). This findspot also lies close to a number of Early Bronze Age cremations and O’Connor suggests that the place was deliberately chosen for the deposition of the axeheads because of the presence of the prehistoric monument and cemetery (2007b, 76). Poolewe is thus currently unique in Scotland for this period; however, its links in terms of composition and depositional practice both echo aspects of preceding Late Bronze Age practices and encompasses traits of the Earliest Iron Age hoards buried further south later on. Previously, the extensive deposition of metal objects in the Late Bronze Age, followed by relatively few bronze objects entering the archaeological record in the Earliest Iron Age had often been taken to represent the collapse of the bronze industry and the wholesale abandonment of bronze (e.g. Burgess 1979; Thomas 1989). These theories are now recognized as inadequate (Needham 1990, 130–140; 2007; O’Connor POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 34 2007a; Cunliffe 2015, 11ff.), and the picture is in fact more nuanced. Changes relating to metal deposition were probably linked with broader socio-cultural shifts and perhaps to the changing role bronze played in society more broadly (Needham 2007). Around this time, one sees the development of a variety of architectural traditions in central and northern Britain, particularly across Scotland, including brochs, hillforts, crannogs and embellished hut structures (Henderson 2007; Pope 2007; Ralston and Ashmore 2007). Several sites now present evidence for occupation spanning the Late Bronze Age–Earliest Iron Age/Early Iron Age (c.1100–400 BC), such as Cladh Hallan, South Uist (Parker Pearson, Sharples and Symonds 2004, 70) and Bellfield Farm, North Kessock (Murray 2012). Furthermore, whilst the object forms that were the staple of the preceding Ewart Park metalworking tradition ceased to be produced and deposited (Needham 2007), new forms of bronze objects emerged during and after this 800 BC threshold, albeit on a limited scale, including the various axehead types mentioned above, as well as Gündlingen swords, chapes, vessels, horse gear and razors (O’Connor 2007a; Gerloff 2010; Boughton 2015). Although its social role changed, bronze continued to be an integral part of Earliest Iron Age social life, alongside the burgeoning industry of iron working. Supporting this, an increasing number of Earliest and Early Iron Age metalworking sites, involving both ferrous and non-ferrous metallurgy, are now known from northern Scotland. At the Iron Age hillfort at Dun Knock, Perth and Kinross, nonferrous metalworking is indicated by the recovery of a crucible fragment from a ditch, as well as a range of ferrous metallurgical debris (Poller 2015, 8–9) and a possible Late Bronze Age phase has been suggested by radiocarbon dates (Daglish et al. 2009, 5; Poller 2015, 24). Evidence for Iron Age ferrous and non-ferrous metalworking is likewise known from Moncreiffe House henge, Perth and Kinross, and High Pasture Cave, Isle of Skye (Stewart 1985; G. Cruickshanks pers. comm.). Although the exact date of metalworking at Moncreiffe House is uncertain, a fragmentary Late Bronze Age chisel suggests it could be earlier in the Iron Age, rather than later. At High Pasture Cave possible non-ferrous metalworking crucibles have been recovered from the earliest Iron Age phases (G. Cruickshanks pers. comm.). To this we can add the evidence of a metalworking site at Rosskeen discovered in 1847 and from which two stone moulds for POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 35 Sompting axeheads were recovered (Wilson 1863, 345–346), as well as the mould fragments from Little Dunagoil. Finally, at North Kessock, fragments of clay moulds for producing at least five socketed axeheads, a spearhead, and a range of small tools, including four sickles, four gouges, a possible awl, a socketed knife and a tanged knife, were found in pits associated with Early Iron Age houses (Cowie and Kruse in Clark et al. 2017, 60). Radiocarbon dates from these pits fell between c.800–400 BC and typologically the moulds included Late Bronze Age object forms and Earliest Iron Age socketed axeheads (Cowie and Kruse in Clark et al. 2017, 60); this suggests that the production event took place towards the beginning of the eighth century BC, broadly contemporary with the deposition of Poolewe. Additionally, over 16kg of metalworking waste was recovered, including slag, furnace lining and other waste material diagnostic of ironworking (Murray 2012, 24). These sites emphasise that the industry of bronze production did not cease with the introduction of iron in Scotland around 800 BC, but rather continued alongside it and thus we can infer changes in depositional practices and a shift in the social role of bronze, rather than the mass abandonment and cessation of production of a useful material (cf. Needham 2007). Moreover, the evidence is dominated by the remains of axehead production and the discoveries at High Pasture Cave, Rosskeen, and North Kessock help contextualise the Poolewe hoard in what appears to be an industrious region of northern Scotland. The metallurgical analysis of Poolewe is important in this regard as it confirms new metal was entering production spheres and that supplies of metal were changing in line with southern Britain and other parts of Europe (Roberts et al. 2015; see Northover above). Multiple impurity patterns confirm that the objects were likely produced from different sources of metal: the three Sompting axeheads, for instance, were produced from three pools of metal though are broadly contemporary typologically. Put another way, this was not a homogenised pool of metal. A similar observation has been noted for the Earliest Iron Age axeheads from Langton Matravers, Dorset (Roberts et al. 2015, 379). This could infer production at different workshops or by craftspeople with access to different sources of metal. If we also consider that some of the objects may be typologically earlier, such as the cup-ended ornament, the distinctive metallurgical composition is certainly conducive to the theory that these objects were accumulated POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 36 from different sources over time. Needham (2007, 54) highlighted the important point that there is little evidence of residual Ewart Park metalwork entering the Llyn Fawr material record, which we might otherwise expect given the high volume of artefacts produced and deposited during the period 1000–800 BC, the Poolewe ornament being an exception. It seems that if any was still in circulation it was either deliberately excluded or recycled; however, the limited metallurgical analyses do not support that recycling of earlier Ewart Park material was undertaken on any grand scale (Rohl and Needham 1998, 110; Needham 2007, 54). The Poolewe metallurgy reinforces this shift towards new supplies and perhaps a break with established metalworking traditions. Regarding the social value of bronze, Needham (2007) has suggested that whilst it served a myriad of economic, political and symbolic functions leading up to 800BC, post800BC there was a significant downturn in the extent to which it was used, circulated and deposited. Although the broader evidence certainly suggests this is true, it is also clear that certain bronze objects, particularly axeheads, continued to function both as utilitarian and symbolic objects. Investigations into axe use have been relatively limited, with the only published experiment involving replica socketed axes being Roberts and Ottaway’s (2003) paper utilising Late Bronze Age axehead forms. Nonetheless, wear-analysis of notches and nicks, sharpening striations and asymmetrical wear infers that many of the axeheads deposited during the Late Bronze Age were well-used tools (Roberts and Ottaway 2003; Turner 2010; Knight 2018, 444). Similar evidence is also observed on some Sompting axeheads (Boughton 2015). However, many axeheads deposited in the Earliest Iron Age show limited signs of wear, often deposited as-cast. The most extreme examples of this are the Portland and Blandford types from Wiltshire and Dorset, which were not only fresh from the mould, but also produced in a high-tin alloy that was extremely brittle and created a silvery surface appearance (Boughton 2015; Roberts et al. 2015). These latter axeheads would have been completely unsuitable for use and it has been suggested that the appearance was designed to imitate the colour of iron (Roberts et al. 2015, 388-389); these axeheads may thus have been purely symbolic. In Scotland, we can turn to proxy evidence for the use of axes during the period 800–600 BC, such as the structural posts from Oakbank crannog at Loch Tay, Perth and Kinross, which bear tools marks that closely compare to the form of Earliest Iron Age axeheads (Sands 2016 POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 37 [1997], Ch.7). The variable evidence of wear on the Poolewe axeheads, with some showing signs of repeated resharpening and others appearing to be unused, suggests that not all axeheads functioned similarly within society. These are generally considered multi-functional tools, though also likely had a symbolic aspect, valued in some way for their form. In light of the evidence for the continued production, circulation, use and deposition of socketed axeheads and other implements in Scotland, we can recognize that the hoarding of objects at Poolewe represents the last vestiges of a social practice that was waning around 800 BC in Scotland, though continued elsewhere in Britain until at least 650-600 BC and was taken up again later in the Iron Age. Why exactly hoarding ceased in Scotland at this time is unclear, though as there was clearly still an industry of bronze production and deposition we may speculate that it could be linked to broader ideological shifts, e.g. hoarding was no longer a mechanism by which communities managed their social, political and economic lives in this area. This lack of hoards is not isolated to Scotland, with similar conspicuous absences in central, north-eastern and south-eastern England, which contrast significant clusters of hoards in central southern Britain, south Wales, East Anglia and Cornwall (see Boughton 2015, fig.6.32). The absence in south-eastern England is particularly striking given the volume of metalwork deposited during the preceding Ewart Park period (1000-800 BC) (O’Connor 2007a, 645). The exact nature of this distribution of practices would repay further attention. At Poolewe, the grouping of diverse objects attests to a practice that was common in Scotland throughout the Late Bronze Age and we can suggest that this hoard may have been deposited by a community that was still hoarding material in accordance with long-established ideas. The deliberate fragmentation of one of the axeheads and the removal of a vessel handle likewise recalls Late Bronze Age practices, and the cupended ornament, an oddity in itself, was possibly in longer circulation than other objects. However, at Poolewe we also see a more Earliest Iron Age perspective on depositing complete, as-cast axeheads with limited signs of use. To this we can add the metallurgical evidence that suggests a shift in supply networks, as well as the production and gathering of objects over time. This hoard thus allows us a glimpse at the selection and accumulation of objects prior to deposition, as it incorporates earlier and later object POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 38 styles, used and unused objects, different and innovative manufacturing techniques, as well as older and newer ideas about how objects should be treated prior to deposition. It is not a homogenous group, conforming to a single way of undertaking a deposit. It may in fact represent the coming together, and indeed transformation, of different people or communities with differing ideas about how to approach the deposition of metalwork during a period when the social value of bronze was changing. Each object may have been contributed for specific individual reasons. The very notion of hoarding and burying these objects in a wetland location indicates the continuing social concepts of appropriate depositional practices, with the radiocarbon date indicating this occurred during the eighth century BC when broader social changes were in full swing. Whilst hoarding and depositing objects continued in parts of southern Britain between 800–600 BC, the lack of comparable deposits in Scotland except for Poolewe becomes stark. Viewed from this perspective, we can suggest the Poolewe hoard may in fact be the last Bronze Age hoard deposited in Scotland. Acknowledgments This article has benefited from the insights of several specialists and we are grateful to Sophie Adams, Barbara Armbruster, Katharina Becker, Mary Cahill, Trevor Cowie, Adam Gwilt, Fraser Hunter, Anna Lewis, Mark Lodwick, Alison Sheridan and Theo Skinner. Derek Hamilton kindly provided comments on the sampling and dating of the material. We are extremely grateful to Jim Buchanan for helping us determine the probable location of the Poolewe hoard and for producing maps and images of the local area. Thanks also to the two anonymous reviewers whose thoughtprovoking comments have enhanced this paper. In particular, we wish to thank Karen Buchanan, Gemma Cruickshanks and Susan Kruse for commenting on earlier drafts of this paper. Any errors are our own. POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 39 Figure Captions Figure 1: The Poolewe hoard. Photo: Neil McLean © National Museums Scotland Figure 2: Probable findspot of the Poolewe hoard. Map (left): courtesy of Jim Buchanan. Figure 3: Oblique drone photograph of findspot of Poolewe hoard. Photo: Jim Buchanan. Figure 4: The Poolewe hoard. Illustration Marion O’Neil © National Museums Scotland Figure 5: The hammer blow on the fragmented socketed axehead. Photo: M. Knight © National Museums Scotland Figure 6: The calibrated date on the wooden axe haft fragment (SUERC-81222) (provided by SUERC, University of Glasgow) Figure 7: The bronze cup-ended ornament. Photo: Neil McLean © National Museums Scotland Figure 8: Comparisons of trace elements of socketed axeheads from select Earliest Iron Age / Llyn Fawr copper alloy hoards in Britain with Poolewe: (a) lead and tin; (b) nickel and antimony. Figure 9: Comparisons of trace elements from objects in Late Bronze Age hoards in Scotland with Poolewe: (a) lead and tin; (b) nickel and antimony; (c) antimony and arsenic. Figure 10: A map of Earliest Iron Age / Llyn Fawr axeheads and moulds from Scotland. Sites mentioned in text are labelled. Finds with an imprecise findspot (e.g. ‘Scotland’) have not been mapped. Table captions Table 1: Compositional analysis of the Poolewe objects Table 2: Radiocarbon dates directly associated with Earliest Iron Age / Llyn Fawr metalwork in Britain Table 3: Certain or possible copper alloy cup-ended ornaments from Britain and Ireland POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 40 Bibliography Anderson, J. 1892. “Notice of the discovery of a hoard of the Bronze Age, consisting chiefly of personal ornaments of bronze, amber and gold, at Balmashanner, near Forfar.” Proceedings of the Society of Antiquaries of Scotland 26 (1891-2): 182– 88. Boughton, D. 2015. “The Early Iron Age Socketed Axes in Britain: A Study of Their Typology, Distribution, Associations and Contexts.” unpublished PhD Thesis. University of Lancashire. Boughton, D. 2019. “Connecting with the Past: Earliest Iron Age Multi-period Hoards in Wessex.” In Objects of the Past in the Past. Investigating the Significance of Earlier Artefacts in Later Contexts, edited by M.G. Knight, D. Boughton and R.E. Wilkinson, 42–59. Oxford: Archaeopress (Access Archaeology). Bradley, R. 1998 [1990] The Passage of Arms 2nd edition. Oxford: Oxbow. Bradley, R. 2000. An Archaeology of Natural Places. London: Routledge. Bronk Ramsey, C. 2009. “Bayesian Analysis of Radiocarbon Dates.” Radiocarbon 51(1): 337–360. Bronk Ramsey, C. 2017. “Methods for Summarizing Radiocarbon Datasets.” Radiocarbon 59(2): 1809–1833. Buchan, W. 1905. “Notes on a Bronze Caldron Found at Hattonknowe, Darnhall, in the County of Peebles.” Proceedings of the Society of Antiquaries of Scotland 39 (1904-1905): 14–20. Burgess, C. 1971. “Some Decorated Socketed Axes in Canon Greenwell’s Collection.” The Yorkshire Archaeological Journal 42: 267–272. Burgess, C. 1979. “A Find from Boyton, Suffolk, and the End of the Bronze Age in Britain and Ireland.” In Bronze Age Hoards. Some Finds Old and New (British Archaeological Reports British Series 67), edited by C. Burgess and D. Coombs, 269–282. Oxford: British Archaeological Reports. Clark, G., T. Cowie and S. Kruse. 2017. Feats of Clay. Bronze Age metalworking around the Moray Firth. Inverness: North Kessock and District Local History Society. POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 41 Clarke, D.V. and M.M.B. Kemp. 1984. “A Hoard of Late Bronze Age Gold Objects From Heights of Brae, Ross and Cromarty District, Highland Region.” Proceedings of the Society of Antiquaries of Scotland 114: 189–198. Coles, J.M. 1960. Scottish Late Bronze Age metalwork: Typology, distributions and chronology. Proceedings of the Society of Antiquaries of Scotland 93 (1959–60): 16–134. Coombs, D.G., J.P. Northover and J. Maskal. 2004. “Tower Hill Axe Hoard.” In Uffington White Horse and Its Landscape: Investigations at White Horse Hill, Uffington, 1989-95 and Tower Hill, Ashbury, 1993-94, by D. Miles, S. Palmer, G. Lock, C. Gosden and A.M. Crummy, 203-224. Oxford: Oxford Archaeology, Thames Valley Landscapes Monograph 18. Cowie, T.G., J.P. Northover and B. O’Connor. 1998. “The St Andrews, Fife Hoard: Context and Chronology in the Scottish Late Bronze Age.” In L’Atelier du bronzier en Europe du XXe au VIIIe siècle avant Notre Ère, III (Session de Dijon), edited by C. Mordant, M. Pernot and V. Rychner, 141–154. Paris: Èditions du CTHS. Cunliffe, B. 2015. “Facing the Northern Ocean: The British Late Bronze and Iron Ages in their continental perspective.” In Scotland in Later Prehistoric Europe, edited by F. Hunter, 5-17. Edinburgh: Society of Antiquaries of Scotland. Curwen, E.C. 1948. “A Bronze cauldron from Sompting, Sussex.” Antiquaries Journal 28: 157–163. Daglish, C., O. Lelong, G. MacGregor, and D. Sneddon. 2009. “Dunknock Hillfort Excavations 2009. Data Structure Report.” unpublished GUARD Report. Available via: https://www.gla.ac.uk/media/Media_183911_smxx.pdf. Donahue, D.J., J.W. Beck, D. Biddulph, G.S. Burr, C. Courtney, P.E. Damon, A.L. Hathway, L. Hewitt, A.J.T. Jull, T. Lange, N. Lifton, R. Maddock, L.R. McHargue, J.M. O’Malley, and L.J. Toolin. 1997. “Status of the NSF-Arizona AMS Laboratory.” Nuclear Instruments and Methods in Physics Research B 123: 51–56. POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 42 Dunbar, E., G.T. Cook, P. Naysmith, B.G. Tripney and S. Xu. 2016. “AMS 14C Dating at the Scottish Universities Environmental Research Centre (SUERC) Radiocarbon Dating Laboratory.” Radiocarbon 58(1): 9–23. doi:10.1017/RDC.2015.2. Eogan, G. 1983. Hoards of the Irish Later Bronze Age. Dublin: University College Dublin. Eogan, G. 1994. The Accomplished Art: Gold and Gold-working in Britain and Ireland during the Bronze Age. Oxford: Oxbow Books. Fontijn, D. 2008. “‘Traders’ Hoards’. Reviewing the Relationship Between Trade and Permanent Deposition: The Case of the Dutch Voorhout Hoard.” In Hoards from the Neolithic to the Metal Ages. Technical and codified practices (British Archaeological Reports International Series 1758), edited by C. Hamon and B. Quilliec, 5–17. Oxford: Archaeopress. Gerloff, S. 2010. Atlantic Cauldrons and Buckets of the Late Bronze and Early Iron Ages in Western Europe. Prähistorische Bronzefunde II.18. Mainz: Franz Steiner Verlag Stuttgart. Gwilt, A., M. Lodwick, J. Deacon, N. Wells, R. Madgwick and T. Young. 2016. “Ephemeral Abundance at Llanmaes: Exploring the Residues and Resonances of an Earliest Iron Age Midden and its Associated Archaeological Context in the Vale of Glamorgan.” In Celtic from the West 3: Atlantic Europe in the Metal Ages: Questions of Shared Language, edited by J.T. Koch and B. Cunliffe, 293– 332. Oxford: Oxbow Books. Hansen, S. 1994. Studien zu den Metalldeponierungen während der älteren Urnenfelderzeit zwischen Rhônetal und Karpatenbecken. Bonn: Universitätsforschungen zur prähistorischen Archäeologie 21. Hansen, S. 2016. “A Short History of Fragments in Hoards of the Bronze Age.” In Materielle Kultur und Identitätim Spannungsfeld Zwischenmediterraner Weltund Mitteleuropa, edited by H. Baitinger, 185–205. Mainz: Römisch-Germanisches Zentralmuseum. Hawkes, C.F.C. and R.R. Clarke. 1963. “Gahlstorf and Caister-on-Sea. Two Finds of Late Bronze Age Irish gold.” In Culture and Environment. Essays in honour of Sir POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 43 Cyril Fox, edited by I.L.L. Foster and L. Alcock, 193–250. London: Routledge and Kegan Paul. Henderson, J.C. 2007. The Atlantic Iron Age: Settlement and Identity in the First Millennium BC. London: Routledge. Hingley, R. 2009. “Esoteric Knowledge? Ancient Bronze Artefacts from Iron Age Contexts.” Proceedings of the Prehistoric Society 75: 143–165. Huth, C. 1997. Westeuropäische Horte der Spätbronzezeit. Regensburg: Regensburger Beiträge zur Prähistorischen Archäologie. Jolly, W. 1880. “Notes on Bronze Weapons and Other Remains Found near Poolewe, Ross-Shire.” Proceedings of the Society of Antiquaries of Scotland 14 (1879-80): 45–49. Joy, J. 2016. Hoards as collections: re-examining the Snettisham Iron Age hoards from the perspective of collecting practice. World Archaeology 48(2): 239–253. doi: 10.1080/00438243.2016.1152197. Knight, M.G. 2017. “The Deliberate Destruction of Late Bronze Age Socketed Axeheads in Cornwall.” Cornish Archaeology 56: 203–24. Knight, M.G. 2018. “The Intentional Destruction and Deposition of Bronze Age Metalwork in South West England.” unpublished PhD Thesis. University of Exeter Knight, M.G. 2019a. “A Radiocarbon Date for a Llyn Fawr Phase Hoard from Scotland.” PAST 91: 13–14. Knight, M.G. 2019b. “Doubtful Associations? Assessing Bronze Age ‘Multi-period’ Hoards from Northern England, Scotland and Wales.” In Objects of the Past in the Past. Investigating the significance of earlier artefacts in later contexts, edited by M.G. Knight, D. Boughton and R.E. Wilkinson, 19–41. Oxford: Archaeopress (Access Archaeology). Knight, M.G. forthcoming. “Accumulations over time. Recognising time-depth in Bronze Age and Iron Age metalwork hoards in Britain.” In Hoarding and Deposition in Europe from later prehistory to the medieval period. Finds in context, edited by E. Durham, I. Bertrand, J. Hall, J. Keily and M.G. Knight. Instrumentum Monographies Series. POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 44 Lawson, A.J. and M.J. Allen. 2000. Potterne 1982-5. Animal Husbandry in Later Prehistoric Wiltshire. Salisbury: Trust for Wessex Archaeology. Liversage, D. 2000. Interpreting Impurity Patterns in Ancient Bronze: Denmark (Nordiske Fortidsminder, Serie C, I). Copenhagen: Det Kongelige Nordiske oldskriftselskab. Lodwick, M. 2010. NMGW-7ECFB8: A BRONZE AGE DRESS FASTENER (UNKNOWN) Available via: https://finds.org.uk/database/artefacts/record/id/289717. Last accessed: 19/03/2020. MacKie, E.W. 1967. “Inveran, Gairloch.” Discovery and Excavation in Scotland 1967: 47. Maryon, H. 1938. “The Technical Methods of the Irish Smiths in the Bronze and Early Iron Ages.” Proceedings of the Royal Irish Academy Section C: Archaeology, Culture, History, Literature 44 (1937/1938): 181–228. McOmish, D., Field, D. and Brown, G. 2010. “The Bronze Age and Early Iron Age Midden Site at East Chisenbury, Wiltshire.” The Wiltshire Archaeological and Natural History Magazine 103: 35–101. Murray, R. 2012. “Bellfield Farm, North Kessock. Area 2 and 3 Archaeological Excavation for Tulloch Homes Ltd.” Unpublished Headland Archaeology Report. Accessible via: https://librarylink.highland.gov.uk/LLFiles/250761/full_250761.pdf. Needham, S.P. 1990. The Petters Late Bronze Age Metalwork. London: British Museum Occasional Paper No. 70. Needham, S.P. 2007. “800BC: The Great Divide.” In The Earlier Iron Age in Britain and the Near Continent, edited by C. Haselgrove and R. Pope, 39–63. Oxford: Oxbow Books. Needham, S.P., C.B. Ramsey, D. Coombs, C. Cartwright and P. Pettitt. 1997. “An Independent Chronology for British Bronze Age Metalwork: The Results of the Oxford Radiocarbon Accelerator Programme.” Archaeological Journal 154: 55– 107. POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 45 Northover, J.P. 1983. “The Exploration of the Long-distance Movement of Bronze in Bronze and Early Iron Age Europe.” Bulletin of the Institute of Archaeology, University of London, 19 (1982/3): 45–72. Northover, J.P. 1987. “Non-ferrous metallurgy.” In Hengistbury Head, Dorset: Volume 1: the prehistoric and Roman settlement, 3500 BC–AD 500, edited by B.W. Cunliffe, 186-195. Oxford: Oxford University Committee for Archaeology Monograph 13. Northover, J.P. 2004. “Analyse der Metallfunde aus Zug-Sumpf, (tr. S. Hämmerle).” In Der Funde der Grabungen 1923-37: Zug Sumpf, by I. Bauer, B. Rückstuhl and J. Speck, Bd 3/1, 102-143 and Bd 3/2 A1-A29. Zug: Kantonales Museum für Urgeschichte Zug. Northover, J.P. 2010. “Metallurgy of Atlantic Cauldrons and Buckets.” In Atlantic cauldrons and buckets of the Late Bronze Age and Early Iron Ages in Western Europe (Prähistorische Bronzefunde II.18), by S. Gerloff, 35-41, 54-55, 61-63, 79-80, 102, 183-184, 189, 251-252, 277-78, 352-365. Stuttgart: Franz Steiner Verlag. Northover, J.P. and N. Wilcox. 2013. “Matthew Boulton’s copper.” In Matthew Boulton: Enterprising Industrialist of the Enlightenment, by K. Quickenden, S. Baggott and M. Dick (eds), 101-110. Aldershot: Ashgate. O’Connor, B. 1980. Cross-Channel Relations in the Later Bronze Age (British Archaeological Reports International Series 91). Oxford: British Archaeological Reports. O’Connor, B. 2007a. “Llyn Fawr Metalwork in Britain: A Review.” In The Earlier Iron Age in Britain and the Near Continent, edited by C. Haselgrove and R. Pope, 64– 79. Oxford: Oxbow Books. O’Connor, B. 2007b. “Two Late Bronze Age Socketed Axes from Tillicoultry, Clackmannanshire, and Similar Lost Axes from Northern Britain.” Tayside and Fife Archaeological Journal 13: 74–79. Palmer, S.C. 1999. Archaeological Excavations in the Arrow Valley, Warwickshire. Birmingham: Birmingham & Warwickshire Archaeological Society Transactions 103. POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 46 Parker Pearson, M., N.M. Sharples and J. Symonds. 2004. South Uist: Archaeology and History of a Hebridean Island. Stroud: Tempus Publishing Ltd. Piggott, S. 1953. “A Late Bronze Age Hoard from Peeblesshire.” Proceedings of the Society of Antiquaries of Scotland 87 (1952-53): 175–186. Poller, T. 2015. “Dun Knock Excavations 2015. Data Structure Report.” unpublished SERF Report. Accessible via: http://www.seriousanimation.com/wpcontent/themes/vislab/serf/pdf/Dun_Knock_Excavation_Report_2015.pdf. Pope, R. 2007. “Ritual and the Roundhouse: A Critique of Recent Ideas on Domestic Space in Later British Prehistory.” In The Earlier Iron Age in Britain and the Near Continent, edited by C. Haselgrove and R. Pope, 204–228. Oxford: Oxbow Books. Ralston, I. and P. Ashmore. 2007. “The Character of Earlier Iron Age Societies in Scotland.” In The Earlier Iron Age in Britain and the Near Continent, edited by C. Haselgrove and R. Pope, 229–247. Oxford: Oxbow Books. Roberts B.W. and B. Ottaway. 2003. “The Use and Significance of Socketed Axes During the Late Bronze Age.” European Journal of Archaeology 6(2): 119–140. Roberts, B.W., D. Boughton, M. Dinwiddy, N. Doshi, A.P. Fitzpatrick, D. Hook, N. Meeks, A. Mongiatti, A. Woodward and P. Woodward. 2015. “Collapsing Commodities or Lavish Offerings? Understanding Large-Scale Metalwork Deposition at Langton Matravers, Dorset.” Oxford Journal of Archaeology 34(4): 365–395. Rohl, B. and S.P. Needham. 1998. The Circulation of Metal in the British Bronze Age: the Application of Lead Isotope Analysis. London: British Museum Occasional Paper 102. Sands, R. 2016 [1997]. Prehistoric Woodworking: Analysis and Interpretation of Bronze and Iron Age Toolmakers. Abingdon: Routledge. Savory, H.N. 1980. Guide Catalogue of the Bronze Age Collections. Cardiff: National Museum of Wales. Slota Jr., P.J., A.J.T. Jull, T.W. Linick and L.J. Toolin. 1987. “Preparation of small Samples for 14C Accelerator Targets by Catalytic Reduction of CO.” Radiocarbon 29: 303–306. POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? 47 Staves, L. 2006. NLM-AC6166: A ROMAN HARNESS FITTING. Available via: https://finds.org.uk/database/artefacts/record/id/127000. Last accessed: 19/03/2020. Stenhouse, M.J. and M.S. Baxter. 1983. “14C Reproducibility: Evidence from Routine Dating of Archaeological Samples.” PACT 8: 147–161. Stewart, M.E.C. 1985. “The Excavation of a Henge, Stone Circles and Metal-working Area at Moncreiffe, Perthshire.” Proceedings of the Society of Antiquaries of Scotland 115: 125–150. Stuiver, M. and P.J. Reimer. 1986. “A Computer Program for Radiocarbon Age Calibration.” Radiocarbon 28: 1022–1030. Taylor, R. 1993. Hoards of the Bronze Age in Southern Britain: Analysis and interpretation (British Archaeological Reports British Series 228). Oxford: Archaeopress. Thomas, R. 1989. “The Bronze-Iron Transition in Southern England.” In The Bronze-Iron Age transition in Europe: aspects of continuity and change in European societies, c. 1200–500BC (British Archaeological Reports International Series S483), edited by M.L. Stig-Sørensen and R. Thomas, 263–286. Oxford: British Archaeological Reports. Turner, L. 2010. A Re-Interpretation of the Later Bronze Age Metalwork Hoards of Essex and Kent (British Archaeological Reports British Series 228). Oxford: John & Erica Hedges. Waddington, K., A. Bayliss, T. Higham, R. Madgwick and N. Sharples. 2019. “Histories of deposition: creating chronologies for the Late Bronze Age-Early Iron Age transition in Southern Britain.” Archaeological Journal 176(1): 84– 133. doi: 10.1080/00665983.2018.1504859 Ward, G.K. and S.R. Wilson. 1978. “Procedures for Comparing and Combining Radiocarbon Age Determinations: A Critique.” Archaeometry 20: 19–32. Watson, P.J. 1999. “Broom, Area E: Bronze Cauldrons.” In Archaeological Excavations in the Arrow Valley, Warwickshire, by S.C. Palmer, 43–50. Birmingham: Birmingham & Warwickshire Archaeological Society Transactions 103. POST-PRINT: Knight et al. Poolewe: The last Bronze Age hoard in Scotland? Wilson, D. 1863. Prehistoric Annals of Scotland, Volume 1 (2nd Ed.). London and Cambridge: Macmillan and Co. Yates, D. and R. Bradley. 2010. “The siting of metalwork hoards in the Bronze Age of south-east England.” The Antiquaries Journal 90: 41-72. 48