As has been referred to in sections 1 and 2, there are at present no indications of either processing or smelting from any of the known early mine sites in the British Isles. Some evidence, however, has been interpreted in terms of processing from Mt. Gabriel in Ireland (O’Brien 1994), consisting of ore-bearing and non-ore-bearing material that has been deliberately sorted into various size components. More recent excavations by O’Brien (pers. comm.) of the Early Bronze Age mines at Ross Island near Killarney have revealed material that could be interpreted as remains of ore roasting or possible smelting. As yet the age of this activity has not been conclusively confirmed, and if the present single C14 date of Dark Age period is to be believed then some re-interpretation will be necessary.
Early mining sites through Europe and the near East contrast markedly with those of Britain and Ireland, in that many have associated evidence of smelting at, or more commonly near to, the mine sites (eg. S.W Spain, Cabrières in France, Mitterberg in Austria, Kestel in Turkey). Observation at these sites seems to suggest that the availability of water for ore-washing and fuel for smelting may be the main factors governing their positions. It is also possible that, as has been suggested by Bick (pers. comm.) for reasons of ritual, early miners would not have allowed smelting at the mine itself. In determining the position of smelting and processing sites in Britain it may therefore be necessary to investigate locations distant from the mines, where the necessary fuel and water for these operations were more available.
The following sections will describe both historical and archaeological information pertaining to potential evidence for processing and smelting at the Great Orme. From both this information, and that from other known sites, it may then be possible to determine where the processing/smelting sites were located and what factors dictated their position. Estimates of ore production and smelted copper metal will also be discussed, which at present are thought to be in the region of a few hundred tonnes.
In view of the large quantities of ore that are likely to have been removed from the Great Orme in the Bronze Age, it is reasonable to assume that the ores must have been at least processed, if not smelted, at or near to the mine. With this in mind a number of documentary sources were studied. One particular article (Roberts 1913) on the wells and springs of the Great Orme and surrounding district mentions three sites where ore tailings or washings were found by miners in the early 19th century. As was common place last century, these tips were regarded to be Roman in origin, based on the argument that ‘modern miners’ always brought water for washing to the mine, while the Romans instead took their ores to the water sources. A reference to another similar site on the western side of the headland was also found (Parry 1863). Details of the four sites (figure 28), together with present day descriptions are given below.
Ffynnon Galchog (‘The Lime Well’) - SH7775 8367
This spring is situated on flat ground on the northern slopes of the headland and lies 0.8km to the north east of the mine. Within this area a number of springs can be seen, all appear to follow a similar contour level (approx.100m AOD) over a distance of at least 350m. The account (Roberts 1913) from earlier this century describes that this well had been used by the Romans as a copper washing place:
“on the flat ground near this well many years ago, scores of tons of
copper slime were carted down the Orme and shipped to the smelting
works. This slime was the washings of the copper which was let to
run on the surface of the mountain and being left for a considerable
period was formed into semi hard material. Distinct traces of this yellow material can be seen to this day. This well was never known to go dry
even on occasions of great droughts”
This description indicates that a large quantity of material with a sufficiently high level of copper existed at this location last century, much of which was subsequently removed. Investigations of the site in recent years confirm there are indeed remnants of the tips as described last century. The tips are now largely grassed over, but can be seen to give rise to a distinct topography which differs from that of the surrounding hillside. Initial studies completed in 1990 concluded that the tip material was of a dolomitic nature in contrast to the local limestone and bore many similarities to spoil deposits from the mine site. Details of these investigations are given in a following section.
Ffynnon Rufeinig (‘The Roman Well’) - SH 7652 8379
Also known as Ffynnon Lletyr Madoc (The Well of Madoc’s Abode) this spring lies on flat ground adjacent to a farm track on the northern slopes of the Orme, 0.9km to the north west of the mine site. Like the previously mentioned site, this spring, as its name implies is considered to have been another Roman ore washing site. Roberts (1913) describes “many tons of copper slime have been carted down the Orme from this spot and shipped, and traces of the yellow stain of the copper washing is visible to-day”. He also comments on the remains of unrelated buildings alongside the tips which he supposes to be of 13th century origin.
A cursory examination of this site confirmed the presence of remnant tips of fine dolomitic spoil similar to that of Ffynnon Galchog, and also similar to material from the mine site. Detailed investigations have yet to be completed, though it is likely these will proceed during the next few years.
Ffynnon Gaseg (‘The Mare’s Well’) - SH 7532 8404
Parry (1863) describes a washing site on flat ground near to the Llech gully on the extreme north west cliffs of the Orme. He argues that it was the early Welsh, rather than Romans who were responsible for carrying the copper ore to open places on the mountain where it was convenient to wash it. It is suggested (Parry 1863) that a small quantity of copper slime was also removed from this site last century.
A preliminary examination of this site which is 1.9km from the mine, indicated an area of grass covered hummocky ground similar in form to the two previous sites. As yet no excavations have been initiated here, but material derived from animal burrows would seem to indicate a component of dolomitic sand sized particles suggestive of mine spoil. This site has been surveyed by Bibby (1979), who interpreted it as a settlement site with hut circles and a possible burial cairn but made no reference to any associations with ore washing.
Ffynnon Tyn y pwll (‘The House at the Pool’) - SH 779 825
This well was formerly situated on flat ground to the south east of the Orme, near to the present Siloh Chapel at the junction of Arvon and Gloddaeth Avenue. The site was built over during the later part of last century and lies about 1km from the mine. From the account given by Roberts (1913) it would seem that this was the largest of the assumed Roman washing sites where “years ago many hundred tons of copper slime was shipped”. He goes on to describe:
“connecting this with the fact that a Roman burial ground existed in
a mound which formerly stood on Madoc Street, where it is believed
some of the Roman mine onlookers were buried, it may be reasonably inferred that a large Roman copper depot at some remote period
occupied the land where the above mentioned chapel is now erected”.
Unfortunately no indications of either the washing site or burial chamber now exist. Interestingly though, guide books to Llandudno last century (eg. Hicklen 1863) also refer to the burial mound and a cromlech that was situated towards the railway station. Structures such as these are more typical of a Neolithic to Early Bronze Age origin, as opposed to the Roman period suggested in the account.
In 1990 permission was gained to proceed with a preliminary investigation of the
Ffynnon Galchog springs and associated assumed ore-washing tips. Initial field walking of the remnant tips enabled the extent of the tips to be estimated. This was decided simply by comparing the topography of the tips to that of the surrounding ground which clearly indicated a number of differences between the two ground forms. The tips were of an irregular hummocky form with thick grass and bracken covering most surfaces while in other areas, particularly the flatter ground between the tips, vegetation was less developed and more straggly. A number of weathered limestone boulders were observed over the site, typically these lay in the hollows between adjacent tips and some displayed alignment to suggest they may have been part of a man made structure (eg. wall). Some of the hollows were quite damp and displayed runnel features indicating increased surface water flow had recently occurred. A few seepages at the same locations indicated that ground water was percolating to surface, probably down the inclined limestone strata that dip to the north. In contrast to the tips, the surrounding ground was markedly flat lying and gently sloped to the north. Vegetation cover was similar to the tip areas but without the patches of stunted growth. The area occupied by the tips was estimated to be 90x60m.
Prior to excavation, three holes were made at different locations by core drilling to the thicker sequences of the tip material. This indicated the tips comprised of silt, sand and some gravel sized particles of dolomitic material, with a small proportion of the larger particles displaying green staining due to malachite. In many respects the material was very similar to the finer grained spoils found at the mine site. During the field search of the site it was noted that none of the surrounding limestone country rock showed any indications of dolomitisation. The complete absence of any mining in the immediate vicinity also confirmed this observation, as copper mineralisation only occurs in those limestones that have been dolomitised. It was therefore concluded that the material constituting the tips had clearly been moved here at some time before being excavated in the 19th century. This prompted trial excavations to proceed.
Two trenches were excavated through one of the more pronounced tips roughly situated towards the south east corner of the site (Lewis 1990c, Jones 1994). Both yielded sequences of graded dolomite silts, sands and medium gravels within which were several copper stained fragments of bone, six spalls from hammerstones and a few fragments of mussel shells. The size and well-sorted nature of the individual horizons in the tips would imply that some form of grading of material had been achieved, with the majority of the deposit being silt to sand size particles (plate 21). Mineralogically the material contained a high proportion of dolomite with lesser amounts of mudstone, limestone, noticeable chert and occasional but obvious fragments of malachite and altered chalcopyrite. At the base of the excavated tip material a charcoal/earth layer was noted, this appeared to be an original ground surface.
The presence of bone and stone tools, malachite, altered chalcopyrite and dolomitic silts, sands and gravels comprising the tips display all the characteristic of similar spoils identified from the mine site which are known to be of Bronze Age origin. It was then speculated that this ‘washing’ site could therefore have been utilised for ore processing in this early period. With this in mind, sampled bone from the excavated tips was submitted for C14 dating to the British Museum. The resulting date, rather curiously indicated a Dark Age period of 1200+/-60BP (BM-2802) which calibrated to 720-740 and 680-960 AD. This apparent anomaly will be discussed in the following section.
In May 1993 a severe storm caused extensive flooding in the Llandudno area. This resulted in a number of land slips over the Great Orme, one of these ‘washouts’ 160m down slope and north of the excavated site (W1, figure 28) revealed a series of stratified deposits in the eroded face of the slope. The exposed material consisted of dolomitic silts, sands and gravels with overlying natural ground of earth and fragmented limestone. In places this sequence was underlain by impressive deposits of tufa that had also formed at the time of increased ground water flow through this small valley. Five spalls from likely hammerstones and four fragments of bone were found associated with these eroded deposits. The dolomitic deposits and the artefacts contained in them bore many similarities to the excavated sequences of material from the Ffynnon Galchog site, potentially extending the area covered by these deposits.
In close proximity to the ‘washout’ a circular stone feature, approximately 5m in diameter and resembling a hut circle was observed. To one side of this structure a trench like feature was noted. From the dampness of the ground here and the obvious water seepages it is likely the site had an association with the collection of water rather than as a habitation. However, because of limited excavation, it is not possible to ascribe any age to the site. Present evidence in the form of iron pipe also exposed by the land slip could suggest this was the water supply to a horse trough which was formerly positioned, down slope on the Marine Drive roadway.
Apart from the evidence mentioned above for potential early ore processing, there are no indications or even suggestions as to where roasting or smelting of the ores was carried out on the Great Orme. This complete lack of smelting evidence, such as burnt or contaminated ground, remnant furnaces or presence of slag, seems to typify the early mining sites of Britain and Ireland, contrasting them with the remainder of western Europe where evidence is more common place. The following section will discuss the apparent lack of smelting evidence at the Great Orme and attempt to explain where such sites could be found.
Collective evidence in the form of historical documentation together with field observation, excavation and survey indicates that mining at the Great Orme in the Bronze Age period was on an unprecedented scale not only in Britain but in Europe. A conservative estimate of produced copper metal from surface and underground during this early period in the region of 205 tonnes has been proposed (section 4.4.5), while a more recent study by Randall (1995) revises this figure to a minimum of 885 tonnes and maximum of 1,769 tonnes. Both sets of data are based on a similar mass of excavated ore bearing material in the region of 40,000 tonnes mined from the opencast and underground workings. At this stage it is difficult, because of lack of comprehensive data, to consider these figures as being realistic. However, future research proposals hope to investigate ore production and concentrations, and also rates of metal recovery to eventually gain a more reasoned idea of the final smelted copper from the Great Orme.
Processing
It can be argued that, because of the nature of the ore and gangue rock, a degree of separation could have been achieved underground. This would be particularly true for the more massive easily extracted ores, as they could be partly trimmed of their waste with minimal effort because of the rotted nature of the surrounding dolomite. As activity progressed to deeper parts of the mine, ores would become gradually more difficult to extract, due mainly to the less rotted nature of the ore-bearing dolomites. Removal of this more resistant material is likely to have necessitated more labour intensive systems of mining, including resort to firesetting and use of stone tools. It is also possible that the ores may have become more disseminated once the more massive and easily worked ores had been removed. A point would therefore be reached where the same degree of hand-sorting was no longer possible and a more efficient technique of separation would be required. At its simplest this may have consisted of just washing the dolomite fines from the lumps and fragments of ore. Where ores were of a disseminated or finer grained nature, processing is likely to have involved water separation techniques relying on the differing densities between the malachite, altered chalcopyrite and the surrounding matrix of dolomitic material.
Considering the volumes of ores that were removed in early times it is not therefore unreasonable to assume that the ‘Roman ore washing sites’ could in fact have originated during the Bronze Age period. Unfortunately, the single and possibly anomalous C14 date so far obtained from the Ffynnon Galchog site does not confirm this. Until other dates are obtained from this and the remaining ‘washing sites’ it is not possible to fully interpret the sites and determine whether they belong to the Bronze Age or to other periods when mining and processing may have occurred on the Orme.
At this point it is worth mentioning the occurrence of chert in the Ffynnon Galchog tips. This mineral has a limited distribution throughout the limestones of the Great Orme, and is confined to the ‘Summit Limestones’ towards the top of the headland. These strata have a limited distribution over the area of known mining and occur in one area only, beneath Bryniau Poethion at the northern end of the ore field. Here these beds must have originally outcropped at surface but now lie beneath a mixture of spoil and overburden thought to date largely to the nineteenth century. One account of mining in this area (Williams 1979) describes how two miners in 1849 found copper ore just beneath the surface; within days the whole area was excavated and a quantity of high grade copper ores was removed, the event becoming known locally as the "Welsh California". This discovery indicates that the ores were in a rotted condition and subsequently were easily worked by a system of bell pits which can still be seen. It is rather inexplicable why it took until 1849 for the Bryniau Poethion ores to be discovered especially as the main phases of 19th century mining occurred between 1820-1845. It is also equally remarkable that Bronze Age miners had not discovered and removed these easily worked near-surface ores. Could it be that during the early period this area saw some use (eg. farming) that prevented mining here. The source for the chert in the Ffynnon Galchog tips therefore needs to be considered. The only other sources of chert are the underground workings beneath Bryniau Poethion, where C14 dated Bronze Age passages (location 5) are known (James 1990). However the extent of early workings here within the cherty beds is very limited and extremely unlikely to have generated the volume of chert encountered in the washing tips. The only other possible explanation is that a considerable quantity of rotted dolomitic ores with associated chert had survived as a capping layer to mineralised ground overlying the opencast workings to the south. If this is so, then these chert-containing ore deposits are likely to have been the first to be exploited in early times.
With regard to the four ‘washing sites’ described above it is interesting to note that the reports describe the “slimes or washings” as containing sufficiently high levels of copper to justify their almost complete removal last century. One possible explanation is that the material may have contained an ore type that was not attractive to the earlier miners, for example unaltered chalcopyrite which is not thought to have been mined to any great extent on the Orme (section 4.2). Another explanation may be that the ore remaining in the tips may have been too fine grained to be separated by early processing techniques. Alternatively it is possible that some form of mineral concentration has occurred which is responsible for the “semi hard material” as described at the Ffynnon Galchog site (Roberts 1913). Such effects are likely to be due to differing pH levels in the tips, which may have come about by the oxidation of chalcopyrite ores as described in section 3.5. Such localised acidification may have caused copper to be mobilised and then redeposited and concentrated in certain areas of the tips. A combination of factors may therefore be responsible for the marked copper content of the tips. Recent research (Bache 1995), as part of a soils investigation on the Great Orme, has confirmed that copper levels in the area of Ffynnon Galchog remain remarkably high despite the removal of these tips last century.
Smelting
To date no evidence or associated deposits for any smelting are known from the Great Orme, as is the case for the remainder of Britain. There is debate (Lewis 1994) as to whether smelting remains will be found on the Orme, the argument being that if large quantities of ore were being extracted, the proportion of charcoal required to fuel the smelting would be considerable. Ratios of charcoal fuel to ore are generally assumed to be in the region of 10-15 :1 (Lewis 1990b), and might reasonably be even higher depending on the smelting technique. Obviously this means a requirement for large quantities of timber, which is unlikely to have been obtained from the Great Orme, partly because of its limited size, but more so since such timber would have also been sought for domestic and possible firesetting uses (Lewis 1993a). Economically, taking the ore to the fuel source would be a more viable proposition than transporting quantities of timber or charcoal up the steep slopes that surround the headland. Smelting sites may therefore be situated some distance away from the mine. If this is the case, such sites would ideally require access by boat for both incoming and outgoing materials. A convenient source of timber for the production of charcoal would also be essential. Sites satisfying these criteria occur within easy reach of the headland, either along the coast east or west, or within shelter of Conwy Bay and estuary some several km away to the south of the Orme. Research within these areas is presently being conducted by members of the Bronze Age Coast Project (Bangor University) and Great Orme Exploration society (Bannerman pers. comm.). If the argument that the majority of the smelting occurred away from the site is to be accepted, then it is still possible that limited smelting may have occurred in the vicinity of the mine, particularly during the Early Bronze Age when metal production is likely to have been on a limited scale. However, because of the known extensive mining through the entire Bronze Age period it is very likely that any early smelting remains would have been destroyed or buried by later mining activity.
Summary
From evidence described in section 4. it is now assumed that the early mines of the Great Orme are the largest known to be of Bronze Age origin in Britain and possibly western Europe. Mines of this scale conceivably produced considerable quantities of ore, which is likely to have required organised systems of processing, smelting and distribution of metal. We can therefore argue that the above described ‘Roman ore washing sites’ may more plausibly have originated in Bronze Age times as ore processing sites associated with the known extensive mining. Unfortunately, because of lacking excavated evidence from the four known ‘washing sites’ it is not possible to conclusively confirm this. Similarly, the complete absence of any material that might hint at possible ore smelting, also raises questions as to whether such activities were carried out on the headland. The need to investigate these aspects of processing and smelting have never been more important than now, especially as present work at the mine site conclusively confirms the presence of prehistoric mining, so emphasising the need to determine how and where the ore was dealt with. A program of research to investigate the ‘ore washing sites’ and the possibility of smelting sites away from the Great Orme is presently being proposed by Bronze Age Coast Project, Bangor University - Great Orme Exploration Society (Bannerman pers. comm.) and University of Sheffield (Wager 1996).
In addition to the processing and metallurgical aspects commented on above, a unique opportunity exists for the examination of the socioeconomic aspects of a Bronze Age presence on the Great Orme and in the surrounding region. A remarkable number of field monuments and artefacts from the prehistoric period are known from the local area. Unfortunately few of these have been investigated, but of those that have been (Bibby 1979), evidence does indicate an occupation during the Bronze Age at Kendrick’s Cave (SH 7800 8284) within 1km of the mine.
4.4.5 Estimated Ore and Copper Production
The surveys of the workings completed as part of the geological study have provided useful information for the determination of the known extent of the prehistoric workings. At present the surveys indicate a virtual labyrinth of workings to vertical depths of 70m and extending laterally into the hillside for 240m, covering an area of at least 24000 m2. In total at least 6km of early mined passages have been recognised. New areas of workings continue to be discovered and it is likely, due to the backfilled nature of the passages, that it will be many years before the true extent of the complete prehistoric mine is realised. Tentative estimates of ore and copper production have been considered over the past few years, these have tended to err on the conservative side so as to ensure that ore to recovered metal ratios are not exaggerated, as has occurred previously (Jackson 1980). There are many problems in trying to arrive at realistic estimates of metal production, these chiefly include the quantity and grade of ore to gangue rock, the efficiency of its removal and the ore recovery rate. The estimates which follow assume that carbonate as well as oxidised chalcopyrite ores were removed from the known extent of the early workings. Two areas are considered as follows.
The Opencast
The true extent of this largely prehistoric feature is at present only partly known as its northern extension remains buried under a mixture of recent as well as prehistoric spoil. Estimates of ore bearing rock removed from the opencast are therefore based on the present extent of the feature, which averages 55m long, 23m wide and 8m deep. This indicates the volume of rock removed from the excavation is in the region of 10120m3. Taking the density of dolomite as 2.8, gives a rock mass of 28336 tonnes. At least six of the main north-south veins pass through the opencast, and from their known width and confining horizon of the Pyllau mudstones it would appear that a considerable quantity of ore bearing material once outcropped here. If an arbitrary recovery rate for copper metal of 0.5% is considered this gives figure of 142 tonnes.
Underground Workings
At least 35 of the north-south and 15 east-west veins are known to have been worked during early times. Summing these gives a total length of passage at around 6Km. If an average cross sectional area of 0.75 m2 is considered this gives an overall volume of 4500 m3, equal to 12600 tonnes of ore bearing dolomite. If we assume a similar recovery rate to that of the opencast ores, 0.5% provides a mass of 63 tonnes copper metal. However it is more likely that the ratio of ore to gangue rock would be higher than that of the surface workings. So perhaps a higher rate of recovery should be entertained, for example at 1% recovery 126 tonnes of metal would be produced.
The total copper production from both opencast and underground workings at 0.5% recovery is therefore 205 tonnes, and for 1% recovery from the underground workings would increases this by 63 tonnes to give a total production of 270 tonnes. These copper production values seem very high when comparing them to the known quantities of Bronze Age artefacts from the British Isles (Northover pers. comm.), and will hopefully provide some basis so that this relationship can be more fully explored by other researchers.