This is the text of a paper submitted as part of the required coursework for the Advanced Certificate in Archaeology at King Alfred's College, Winchester. Minimal changes have been made to adapt it for electronic publication.
This essay describes the evidence for the extraction, working and distribution, in the Roman period, of those stone products that are peculiar to the Isle of Purbeck, namely Purbeck limestone (especially Purbeck Marble) and Kimmeridge shale. `Purbeck' in this paper means the part of Dorset south of Poole Harbour and the Frome, and east of Moreton and Owermoigne.
The map of Roman Britain (0rdnance Survey 1991) shows a crowd of sites in Purbeck marked `shale', `stone' and `salt', and several pottery kiln sites. This concentration of industrial finds has been supposed to be due to the unusual enthusiasm of Dorset archaeologists, but is more likely to be real. There is no doubt about the importance of the Dorset black-burnished pottery industry; the shallow margins of Poole Harbour are ideal for salt production; and Purbeck is particularly rich in stone resources, having the Portland freestone, the Purbeck shelly limestone, and the Kimmeridge shale, the last two of which are peculiar to this area. The Portland stone is commercially important but not confined to Purbeck (RCHM, 1952, p. 523; Williams, 1971, p. 179).
In Purbeck there are strata from middle Jurassic to lower Tertiary. The Purbeck rocks dip northwards, that is, the strata are lower on the north than the south. Two bands of harder rock form east-west ridges, the northern of chalk, and the southern of the Purbeck limestone itself, which is counted as topmost Jurassic (British Geological Survey 1976). (No longer.) Below, and so south of, the Purbeck limestone lies the Portland limestone. Below this are the shales and clays of Kimmeridge, exposed only at the coast.
The Purbeck limestone has been worked intermittently from the Roman period to the present day, and apparently also in the Bronze Age (Calkin, 1959a). 24 distinct beds are recognised by quarrymen (Coach House Museum, Langton Matravers---museum of the Purbeck quarry industry). These are distinguished from each other by their characteristic fossil mollusc species. (And in other ways; see this note.)
The topmost bed of the Purbeck limestone is called `Purbeck Marble' (henceforth PM) because it takes a high polish, though it is not a true metamorphic marble. Below PM is the `burr' stone (bed 2), which has been valued decoratively and as an abrasive material. Lower layers have been exploited mostly as common building-stone and as roof-slabs.
Below the Portland stone, the Kimmeridge shale is oil-bearing and has been exploited for fuel from at least 1575 AD to the present. We are mainly concerned with a 60cm seam of shale found only near sea-level at Kimmeridge and Brandy Bays; this `blackstone' was used in Iron Age and Roman times for carved and turned work.
Purbeck Marble is distinguished by its characteristic fossil, the freshwater snail Paludina carinifera. The whole of the PM bed is filled with a dense mass of these shells, each about 6--8 millimetres across. The stone is characteristically bluish-grey in colour, though some sections are greenish or reddish.
PM is most easily confused with the bed below, the `burr', which like PM was used to make mortaria in Roman times. Burr-stone however has a different characteristic fossil. Correction (March 2001)
PM does not weather well. Even in protected situations, its surface loses its polish over time and becomes pitted.
The outcrop of PM all along the north side of the limestone ridge is marked with old workings, but most are modern or mediaeval; no Roman workings have been positively identified, and they will probably have been obliterated by later quarrying. However a number of conjectures have been made about Roman quarry-sites. In my view the most probable sites are at Wilkswood, Dunshay Manor and Worbarrow.
As already mentioned, there is a lack of finished, polished PM on Roman sites in Purbeck. Taken with the good evidence for processing of PM and other stones at building-sites, this indicates that PM for architectural use was generally transported in a rough state.
At Fishbourne, there were regular masons' yards. Stone, including PM, was brought to the site in blocks, tooled into regular shapes with hammers and chisels, then sawn into sheets from 1/8 to 3/4 inch thick. Part-sawn blocks with parallel cuts were found, suggesting use of a multiple-bladed saw as described by Pliny (Correction, Jan. 2004); the iron blades were not toothed but sand and water was used as an abrasive, and much sand was found, assumed to be residue from this process. The polishing was done with stone rubbers (presumably graded; modern polishing machinery uses at least five grades of abrasive discs); the sheets cut to size with a single-blade saw and finally cracked along the cut, leaving a burr to be smoothed off.
One area of the yard at Fishbourne seems to have specialised in cutting small shapes for fine inlay, probably for furniture, while another area produced Purbeck marble mortars and pestles (Cunliffe, 1974).
This was of course on the site of a large building where a great quantity of stone was used. There is evidence however that sometimes stone was at least rough-processed before it stone left Purbeck. At the Norden site just north of Corfe Castle, (Beavis, 1970; Sunter and Woodward, 1987), there is evidence of several kinds of industry, including PM working, from the first to fourth centuries. Slabs were sawn, inlays and tesserae were made, probably mortaria made in PM and other local stones, and PM fragments have been found in a dump probably of the fourth century. Nevertheless evidence of final polishing is missing.
Norden lies just north of the Corfe Castle gap in the northern or chalk ridge. It seems that stone quarried on the southern, limestone, ridge was transported northward through the Corfe Castle gap, roughly prepared in the Norden area, and then sent eastwards to a port on Poole Harbour, maybe at Hamworthy, where there was an important Roman installation in the Conquest and post-Conquest periods.
PM was used in the Roman period for columns and other massive items, mouldings, veneers, inlays, flooring materials, and inscriptions. In excavation reports, fragments of veneer and inlay are the most frequent. This may not reflect the true frequency of use; in particular, tesserae of PM may not always be identified as such.
Besides these `architectural' uses, PM was used to make grinding-bowls (mortaria) and a few ornaments and other small portable articles.
Clarke (1982, p. 211) lists instances of `marble' in British villas, and a similar and complementary list is given as an appendix in Scott (1988). These are very valuable sources, but the original excavation reports sometimes give the origin of the marble and sometimes not, and in this essay I have mostly ignored those that are not explicitly described as PM; a fuller investigation of the use of PM would demand further enquiries about the marbles of unspecified origin.
Remarkably few items of finished Roman PM are known from Dorset. The best known is the tombstone of `Carinus, civis Romanus' (RIB 188), now in Fordington church, Dorchester. The lettering is in 1st-century style (Beavis, 1970; RCHM, 1952, p. 574). A PM column capital comes from the Jordan Hill `temple' near Weymouth, now in the Dorset County Museum. This site has also been interpreted as a marine signal-station; coins and pottery indicate occupation principally in the 4th to 5th century (Drew, 1931 and 1932; RCHM, 1952, p. 616).
At Studland, grid reference SZ 036 825, two cists of weathered PM slabs were found a few yards apart (Calkin, 1952; Farrar, 1955). One contained a decapitated female burial; they have been linked with a nearby Romano-British settlement at Woodhouse Hill. Similar cist-burials in other Purbeck stone were found at Blashenwell, SY 9518 8047 (RCHM 1952, p. 599).
PM tesserae have been found near Dorchester and near Weymouth (see below).
Outside Dorset, the geographical distribution of PM items reaches to London and Chester. Correction (June 2001)
Inscribed PM stones (Collingwood and Wright 1965, Beavis 1970) have been found at Caerleon (1), Chester (1), Chichester (3), Colchester (8), Dorchester (1) (Carinus; the only PM inscription within 50 miles of Purbeck), London (4), Silchester (13), St Albans (7).
The Caerleon and Chester inscriptions are both associated with Legio II Augusta, which must have been at Lake near Poole in the conquest period, and a piece of weathered PM was found in a Lake posthole dated 55--60AD (Putnam 1984). Very likely the legion had something to do with the early exploitation of PM. Further detail (June 2001)
The very large number at Silchester probably reflects the completeness of the excavations, and the number at London by contrast is depressed by the difficulty of excavation. At the least it can be said that there is a wide distribution over the south of the province. Many of the inscriptions are not dated, but those that are are first-century.
Besides the 4th--5th century Jordan Hill capital, fourth-century PM columns have been found at Colchester (Hull, 1958, p. 96). Other columns of `PM or similar' were also found at Colchester (Insula 9). Exeter museum has a column-head which may be PM, but is suspected of being an 18th cent. copy of Roman work (Beavis, 1970).
Paving-stones of PM are known from the Temple of Claudius site at Colchester (Hull, 1958, p. 188--9) and floor-tiles from Fishbourne (Cunliffe, 1962 and 1963). The Colchester stones are presumably before 60 AD.
Tesserae of PM were found at Dorchester (Vidler, 1928; RCHM, 1952, p. 570) and dated to the fourth century. Other possible PM tesserae are recorded at Preston near Weymouth, and in Silchester. It is possible (indeed likely) that other PM tesserae have been found and not recorded as such.
Notable finds have been made as far away as Gloucestershire, Huntingdonshire and Essex. Where datable, they are usually early in the Roman period.
Opus sectile including Purbeck marble was found at the Fishbourne palace (Cunliffe, 1963) and at the villas at Fingringhoe, Essex, and Woodchester, Glos. (Clarke, 1982; Liversidge in Rivet, 1969; Scott, 1988). The Woodchester material is not clearly dated, and the others are first or second century. Much fragmentary PM veneer is known from Colchester (Hull, 1958); Silchester (Boon, 1974); Gloucester (Beavis, 1970). At Godmanchester (Hunts.) the cold bath of the mansio, demolished early fourth century, was lined with vertical PM slabs (Rankov, 1982, p. 363).
Mouldings in PM were found at Fishbourne (Cunliffe, 1962, 1963, 1974); and at Silchester, Colchester and Cirencester (Beavis 1970). The latter are undated; the Fishbourne ones are first or second-century. Fishbourne is noted for lavish use of polished stone, and apparently there were impressive door-casings or architraves in a combination of Purbeck and Turkish marbles.
Other fragments of PM are known from many sites across southern Britain.
Dunning (1949) believed that the PM quarries went out of production towards the end of the second century, but there was a resumption after 350 AD; this view seems to be based on the datable mortaria and the shortage (noted above) of definitely late architectural PM. Beavis (1970) doubts this, noting PM working in later contexts, especially at Norden, but Clarke (1982) cannot point to a late Roman building using PM. The Jordan Hill capital (RCHM, 1952) and the Colchester columns (Hull, 1958) are placed in late contexts and, unless reused from earlier buildings, suggest at least a revival in the PM industry. It is not possible to say how much the industry continued through the third century; maybe it continued in a small way to supply local needs.
Dunning, in Cunliffe (1968) p. 100--111, discusses the distribution of Purbeck mortaria. Some of the items he refers to are not of PM but of similar shelly limestone, such as burr-stone. The typical stone mortar has a curved profile and 4 lugs around the rim, one of which has a pouring-groove cut into its upper surface. Most are dated to the late first century or the second; all others are after 350 AD, and are smaller and straightsided. The geographical distribution reaches to Hadrian's Wall.
Clearly Purbeck mortaria were traded over a wide part of the province, apparently even further than the shale articles discussed below; this is a little surprising since presumably the Purbeck stone was selected for its abrasive properties, but stone with similar properties is available elsewhere in Britain. The gap in the third and early fourth century needs explanation, and Dunning maintains that it indicates a break in the industry; though Beavis doubts this.
As noted above, production of mortaria was sometimes carried on at sites where PM was being prepared for use in local buildings. If we consider mortaria as a by-product of the building industry, we may try to explain the break in the series not by lack of demand for such vessels, but by a lull in building activities requiring PM: during such a period, domestic users would have to turn to other kinds of mortaria, such as pottery ones, while when architectural PM usage was high, there would be a surplus of mortaria which, when finished, could be transported far from the building-sites.
Other portable objects of PM include (Beavis, 1970) L-shaped pestles from Silchester(Boon, 1974); basins from Fishbourne and Silchester (Fishbourne and Reading museums); a dish from Rockbourne Farm which is not a mortarium; a shallow vase from Winterbourne Kingston (Mansell-Pleydell, 1890), a candelabrum foot from Gloucester (Gloucester Museum), rectangular with lion's paw carving; and an `oculist's stamp' from Colchester (Warren, 1944).
At Exeter legionary fortress, the University of Exeter excavations found portable legionary standards of Purbeck marble, which fit in with the theory that the army had a part in the early exploitation of PM. Correction (1999)
The pre-eminence of PM as a high-class decorative material has led to a neglect of the other beds of the Purbeck limestone series by archaeologists. Dunning (1949) called for a study of Roman quarries in general by a geologist, and Williams (1968, 1971) has made an important study of Roman building stone, including a list of known quarries; but there is no specific study of the distribution of Purbeck limestone as building material.
Unsurprisingly, Purbeck stone was used in Roman times for local building (Coach House Museum); the interesting question is how much it was used non-locally. Williams' work is fundamental, and I have tried to supplement it by Scott's catalogue of villa-sites (1988). The main limitation of these studies is that most of the excavation reports on which they rely do not state the place of quarrying of the stone objects. Some of the `limestone' reported could in fact be of Purbeck origin. Further study of the Purbeck quarry industry would have to follow up many possible and dubious records of Portland and other limestone, wherever found, in case it was of Purbeck origin.
One class of object is however distinctive enough to be traceable through these sources. Roofing-slabs of Purbeck limestone are common on buildings currently standing in Dorset, and were also used in the Roman period (Putnam, 1984). Such tiles were normally lozenge-shaped with a single nail-hole, located near the uppermost corner. A variant is pentagonal, the upper corner being cut off by a fifth short side; sometimes two nail-holes are found. A good example can be seen in the Coach House Museum, Langton Matravers. A roof of this material was over the stokehole at Colliton Park (Williams 1971). Correction (July 2001)
Williams (1971, appendix II) lists roofing materials from villas, towns and military sites, identifying some as `Purbeck limestone'. Even after adding a few sites mentioned by Scott (1988), the distribution of Purbeck roofing material does not go north of Wiltshire or east of Hampshire, with a solitary exception at Colchester, Essex. The Colchester example is second century. Most other non-Dorset instances are from the third or fourth centuries, when there was much villa-development especially in Wiltshire and Somerset. Note (June 2001)
Note (February 2001)
Shale can be confused with similar materials, such as jet, cannel-coal, and lignite, which have been used for similar artefacts. Ways of distinguishing between them are described by Thurnam (1871), Williams (1974) and Lawson (1976).
Lignite includes carbonaceous deposits derived from wood, intermediate, both in density and age, between peat and bituminous coal; cannel-coal and jet are forms of lignite. Most lignites are too brittle to be worked; jet is an exception. Jet easily takes a static electrical charge, and breaks with a conchoidal fracture, like glass. The main source of jet in Britain is at Robin Hood's Bay, near Whitby (Northumbria), with small amounts at Cromer (Norfolk) and on the Severn shore at Watchet (Somerset) and Aberthaw (Wales).
Shale is a laminated clayey deposit laid down in water, and so may contain fossil shells. Many shales are found in Britain but only the `Kimmeridge Coal' has been extracted in any quantity. It is appreciably heavier than jet and the other lignites, burns less cleanly and contains oil. Greyish and hard when dug, when wax-polished it becomes black and shiny like jet. In ancient artefacts the oil has been replaced by water, and unless kept constantly wet the shale will delaminate; it must be conserved by replacing the water with an emulsion of wax.
The only exposures of workable-quality shale are on the east side of Kimmeridge Bay (site of quarrying in the 16th to 19th centuries) and at Brandy Bay further west. On the eastern exposure, the outcrops have been affected by coastal erosion and by modern quarrying, and no trace of Roman quarrying can be expected to remain. The western exposure is under-explored, because since 1943 it has been included in the Army ranges. A major shale-working site is known next to the shore at Kimmeridge Bay (Gunder, 1994).
Débris from shale-working is found at many Roman and Iron Age sites in Dorset, but rarely outside the county. In both pre-Roman and Roman periods, most shale was exported from Dorset as finished articles.
The débris consists of lathe-cores (`coal-money'), chippings, and the flint tools used to work the shale, which have a distinctive narrow chisel-like form. Experimental work in recent times has shown that flint tools are preferable to metal for working shale (Calkin, 1953). Some experimental turning with a pole-lathe continues at Upton Park (information from Poole Museum).
There was an early Iron Age industry in which roughouts were chiselled, ground and polished, but no lathe used. The lathe-using industry was continuous from late Iron Age to the fourth century AD (Davies, 1936; Calkin, 1948 and 1953.)
The lathe was used for making bowls, platters, tabletops, spindle-whorls, but especially armlets. A large armlet and then a small one could be cut from the circumference of a single blank, which was then discarded (Calkin, 1973). The discarded cores have holes or sockets for securing them in the lathe; Mansell-Pleydell (1892) describes the various patterns of holes. Cores could be reused as spindle-whorls, or as fuel (Hearne and Smith, 1991); perhaps the Worbarrow collection of cores under a stone cover, referred to by Mansell-Pleydell, was a fuel-store.
Though shale was described as `false jet' by Roman authors (Thurnam, 1871), it was not always regarded as an inferior substitute for jet, at least in Britain, for the shale industry had been active for a long time before a parallel industry in British jet was revived in the first century AD. Jet products were exported to the Rhineland, but a British shale article has been found at Boulogne (Bretz-Mahler, 1971). Besides armlets, jet and shale were used for spindle-whorls, distaffs, beads, necklaces, pendants, finger-rings, hairpins, inlays, furniture, statues (Lawson, 1976).
Small shale objects are widely distributed in Dorset and reached Winchester, Silchester, London, Northampton, Glastonbury, Meare.
The armlet industry was continuous from pre-Roman to Roman times. Pre-Roman armlets in large numbers were found at Glastonbury, Meare, All Cannings Cross (Wilts), and Maiden Castle. Two armlets have been found actually worn by buried people, at Fordington, Dorset (Mansell-Pleydell (1892), p. 182) and at Maiden Castle (Wheeler, 1943, p. 317).
Many lathe-turned armlets are plain, but in all periods it was common to add notches, oblique grooves, segmentation, etc. implying finishing off the lathe. It is difficult to date armlets by their decorative style, though Davies, in Wheeler (1943) attempted to do so; there is a tendency towards a lighter style in the Roman period, perhaps due to improved technique.
The most notable Roman assemblages of shale armlets are 77 items from Silchester (with 19 of jet) (Lawson, 1976) and 88 from Colliton Park, Dorchester (Calkin, 1973), dated to the third and fourth centuries. Smaller assemblages are from Stockton, Wiltshire (21 armlets), and Woodcuts, Dorset (18). More recent work at Colliton Park has turned up an apparent Christian symbol on a shale ornament (Aitken, 1982).
Several furniture-legs of Kimmeridge shale are known. The characteristic pattern has a bowed shape, with an animal-head carved on the outside of the convex bend (Liversidge, in Rivet 1969). The head is sometimes described as `feline' and sometimes as a `griffin'; it is a traditional stylised pattern which appears also on non-shale articles from the Continent, such as a three-legged table from Pompeii and many pictorial representations of furniture.
The best example of such a leg in shale is from Colliton Park, Dorchester, now in Dorset County Museum, and stands 47cm high, suitable for a chair or low table. Others are described in a note.
Biddle (1967) reviewed the incidence of these articles, which are rarely datable, but where found in a datable context it is 2nd century AD or late 1st. The reasonably complete ones are 3 from Jordan Hill, near Weymouth, one fom Rotherly in Cranborne Chase, and the one discovered by Biddle at Winchester. The remainder are fragmentary and mostly from Dorset, except for some from Silchester and London.
Some of these fragments may actually be from larger articles than trays, such as table- or sideboard-tops, or even wall-panelling; Liversidge, in Rivet (1969), refers to `oblong pieces of Kimmeridge shale, sometimes described as trays or wall-veneers, coming mostly from Dorset'. At Dewlish villa, Dorset, the plunge-bath was lined with sheets of shale (Putnam 1984).
Two shale cists have been reported but both are lost; one is said by Mansell-Pleydell (1892) to have been found in the early 19th century at Kimmeridge, containing coarse pottery and an ox-skull; the other was found in 1908 at Worbarrow Bay and recorded at Dorset County Museum (RCHM 1952, p. 612).
Many things besides armlets might be produced on a lathe. Shale vessels were common: fragments of 20 bowls or platters were found at Silchester. Many pre-Roman parallels are known, from Maiden Castle, Sutton Walls (Herefordshire) (Kenyon 1954), etc.; the designs got simpler after the Roman conquest (Lawson, 1976).
Circular trays were also found at Silchester, and some have been classified as table-tops, which are paralleled by circular fragments in the London Museum. Table-tops can be distinguished from trays by the mortises for the legs.
A cylindrical jewel-box of shale, such as might be used on a dressing-table, is included in the Thetford treasure (British Museum; Johns and Potter 1983). This was buried in the late fourth century AD, and is one of the few known shale vessels with a lid; another is an Iron Age bowl from Hengistbury (Bushe-Fox, 1915).
Three curious objects from Silchester, and one from Woodcuts, are shaped like a pulley, or two small cups base-to-base. A hole is bored through the axis. The greatest diameter is 34mm and the axial length around 20mm. (My own suggestion is that they are reels, for string or thread; not toys (yo-yos), as in that case the hole would be functionless.)
Kimmeridge shale is also said to be used for black or grey tesserae in mosaic pavements. Since black tesserae are a basic requirement of mosaic everywhere, it would be interesting to study what different materials were used for this purpose in different parts of the Empire, and in particular what proportion of the black/grey tesserae in Britain were of shale.
We do not know why the industry did not outlast the Roman period. The demand for furniture articles understandably ceased, but why personal ornaments in shale went out of fashion is unexplained. Maybe the industry, being localised in Dorset but exporting to most of Britain and beyond, had become dependent on the cash economy, and fell with it.
These industries show us bits of information about occupations, about tools and technology, about dress and personal ornaments, about domestic furniture and decoration, about trade and transport, and maybe about toys and games, in the Roman period.
Shale-working was the earlier industry; it was well established in Dorset in the Iron Age and for most of the Roman period produced personal ornaments, trays, vessels and platters, and articles of furniture. Although these articles were traded across the whole of Britain, and even abroad, there is little evidence of their manufacture outside Dorset.
By contrast to shale, there is no evidence (apart from an unique Bronze Age cist) of use of Purbeck marble before the Roman conquest. The Romans brought traditions of building in stone and making inscriptions on stone, and were accustomed to using polishable marbles from Italy and elsewhere. Purbeck marble was used, especially in the first and second centuries, alongside imported marbles. The speed with which it was identified and first exploited is remarkable, and suggests that some knowledge of it already existed before the Conquest, which was taken up by the Roman army. In the later Roman period there is relatively little evidence for its architectural use, though foreign marbles are often used. If this decline is real, it may be due to its poor wearing qualities. The late usages at Jordan Hill and Colchester may be reuses of items made earlier, but the poor wearing qualities would seem to make this unlikely, and I would prefer to believe that there was at least some quarrying and shaping of architectural PM in the fourth century.
Architectural PM was usually finished on-site or at least near to the site of use, and I conjecture that mortaria of PM may have become common because they were an easy by-product of the working of architectural PM. With the decline of PM as a building stone, domestic users had to turn to pottery mortaria, though in the 4th century, Purbeck stone mortaria of a different pattern were reintroduced; this industry may have been more centralised. Burr-stone was used as well as PM for mortaria.
Purbeck limestone other than PM was used for building locally, and for roofing-material in Dorset, Wiltshire and Hampshire, especially in the third and fourth centuries. As it was of less value than PM it was usually uneconomic to transport it greater distances, but there is one instance of its use in Colchester.
There are many unsolved problems about Roman Purbeck stone.
Thanks for the help of Mrs Joan Harrison, Langton Matravers museum; Mrs Lander, Lander's Quarries, Langton Matravers; Dr Eleanor Scott, King Alfred's College; Dr Peter Woodward, Assistant Curator, Dorset County Museum; and Miss Mary Spencer Watson, Dunshay Manor.
These are the most serious candidate sites, all of which show some evidence for Roman industrial activity:
Eleven such mortars were found at Richborough, Kent. Others were found at London, Silchester, Colchester, Verulamium, Caerwent, Caerleon, Brading (Isle of Wight), Carisbrooke (Isle of Wight), Lullingstone (Kent), Rotherley (Cranborne Chase, N. Dorset), Woodcuts (Dorset), Wroxeter (one only), Corbridge (one only). (Cunliffe, 1968, p. 100 ff.)
Other fragments of PM are known from Rivenhall, Essex (Clarke, 1982); from Bledlow-cum-Saunderton, Bucks (Scott, 1988, site BU6); and of course from Fishbourne. Dunning's distribution map in the Dorset County Museum (Beavis 1970, RCHM 1952) also records architectural uses of PM at Angmering, Ashtead, Beckley, Bradwell, Canterbury, Darenth, Lockleys, Lullingstone, Park Street, Richborough, Saunderton, and Wroxeter.
There is a notable lack of identified PM in Hampshire (Scott 1988).
Many older excavation reports, and some more recent review articles, refer to items that may possibly be of PM but are not clearly identified as such. Among these are Liversidge, in Rivet (1969), and several of Scott's villa-sites (1988) especially Aldington, Kent (Scott's KE2), Folkestone, Kent (KE34), Ashtead, Surrey (SY3), Hackleton, Northants (NH54), Daventry, Northants (NH30), Weeting, Norfolk (NF193), and Gayton Thorpe, Norfolk (NF62).
PM itself is now worked only occasionally, and permanent quarries are not permitted in the PM outcrop area. At the present time the `thornback', bed 9, is polished and used for ornaments and memorials. (information from Lander's Quarries).
With regard to the durability of PM, observe the 20th century war memorial in Studland parish church; the lower panel, from 1945, is still in good polished condition, but the upper, from 1919, has noticeably deteriorated. The massive Norman font in Wimborne Minster shows a later stage of the process. None of these have ever spent any length of time out of doors.
Shale-working sites in Dorset include Rope Lake Hole, Norden, Ower (Sunter and Woodward, 1987), Hamworthy, Green Island (1st cent. BC) (information from Poole Museum), Brenscombe near Corfe Castle (Scott, 1988, DO8; RCHM, 1952, p. 598; Branigan, 1976, p. 84), Wilkswood (RCHM, 1952, p. 602; Calkin, 1953 and 1959b; also a possible Purbeck Marble quarry site), Blashenwell (RCHM, 1952, p. 599; Frend, 1936), Worbarrow Bay (RCHM, 1952, p. 612), Kimmeridge Bay itself (Gunder, 1994; active 1st cent. AD), Worgret (Hearne and Smith, 1991, p. 93, late Iron Age), Wytch Farm (Cox and Hearne, 1991).
Cores were found at Glastonbury in a pre-Roman context (Bulleid and Gray, 1911) and at Silchester (Boon, 1974), but few in proportion to the quantity of finished armlets. The characteristic flint tools were however found at Dinas Powys in South Wales (Alcock, 1963).
Besides the Colliton leg, other Dorset examples, more or less fragmentary, have been found at Frampton, Preston near Weymouth, and Norden near Corfe Castle. Outside Dorset, similar legs are known from Silchester, Upper Langridge Farm near Bath, Rothley (Leicestershire), Foscott (Buckinghamshire), and Caerleon (Beavis 1970). A fairly general distribution across the southern part of the province is indicated, with a definite concentration about the area of production.
One lathe-turned leg, from Silchester, is also known (Boon, 1974, p. 222, fig. 21; Lawson, 1976, fig. 14) and from its size probably belonged to a stool or couch rather than to a table. Two turned objects from Silchester were identified as bases for legs.