The Leckie broch, Stirlingshire:

Transcription

1 The Leckie broch, Stirlingshire: an interim report Euan W MacKie, BA, PhD, FSA Summary Some ofthe results ofthe excavation ofthe broch at Leckie between 1970 and 1978 are briefly described. The clear stratigraphy and the many closely dateable Romanfinds allow new conclusions to be drawn about the date ofsome Iron Age artefacts and about the origin of the southern Scottish brochs. For example it seems now that some of these brochs were built, presumably with the approval of the Roman Army, during the Flavian period (AD 80100) and probably soon after Agricola's recall. The richness ofthefinds at Leckie, and the many valuable imported items, support the view that the southern broch builders were originally allies ofrome. Later, soon after the start of the Antonine period, things became different and this broch, probably with others, was destroyed possibly by the Roman forces. Introduction The remains of the Iron Age broch at Leckie lie on the southern foothills of 11 km W of Stirling the Forth valley (NGR NS/693940); they stand on a steepsided promontory formed by two streams which have cut deeply into the red sandstone bedrock and which join just below and north of the site before flowing out on to the flat land of the valley bottom to join the River Forth (fig. 1). Until the beginning of the nineteenth century this flat ground on both sides of the river was a vast, almost impassable swamp of peat bogs and pools of water; the Fords of Frew, a few km west of Leckie, show where one passage across to Perthshire was possible, and the other was at Stirling where the valley narrows sharply. The single Roman road which ran north from the fort at Camelon to link up with the camps and forts in Perthshire, Angus and regions further to the northeast ran close to Stirling. The Iron Age strongholds in the western part of the Forth Valley, of which Leckie was one, thus lay at times close to a main Roman route; they also commanded a narrow tract of cultivable land between the boggy valley floor to the north and the Gargunnock Hills to the south. More important, they lay near the only convenient passage from Central Scotland into the northeastern was drained. regions before the Forth Valley The bridge at Stirling was vital to armies in later times as can be seen from the campaigns of the English and Scottish forces in the fourteenth and fifteenth centuries. The discovery by air photography of forts near the Lake of Menteith shows that the Romans brought this whole area under control for a time. The Iron Age stronghold at Leckie was discovered in the late 1960s and excavated between 1970 and 1978 thanks to the kindness of the proprietor, Lord Younger of Leckie (fig. 2). It revealed a clear sequence ofstructures and layers from which it was possible to reconstruct in considerable detail the history of the occupation of thesite. Several kinds ofdating evidenceincluding Roman pottery, glass and coins and radiocarbon measurements combine well with the stratigraphical sequence (Table 1). They give us a vivid picture of the fortunes of the stronghold of one or two wealthy native families during those periods of the first and second centuries AD when the army of Rome was attempting to establish the northern frontier of the Empire in Scotland. The number and quality ofthe finds, and their clear contexts, make possible an unusually precise date for the construction of this particular site, and this helps with the understanding of the origin of the south Scottish brochs. The finds also give us a detailed and vivid picture of the daily life of a native community in central Scotland in the second century AD. The history of the site The origin of the site and its subsequent development are described in outline. The detailed evidence for the bald statements made here will be presented in the final report. 60

2 Phase 1: prebroch: There was a period of occupation on the sandstone promontory before the broch was built. A large number ofpost and stakeholes were found in the interior, cut into either the underlying sandstone or the sandy subsoil which lay on top of it in places. Most of these holes were full of the same black earth which composed the primary broch occupation floor, but a few contained light brown soil; these must therefore have been dug and refilled before any of the broch floor accumulated and they seem likely to antedate the tower entirely. Several of these early postholes form an arc of a circle, and may be the remains of a wooden hut about 5.6 m in diameter the centre of which was well away from the centre of the broch court (fig. 4). No artefacts were found in these early postholes but a group ofsheep bones in one gave a C14 date of AD 80 ± 70 (GU 1370), equivalent to a time span in calendar years centred on about AD 100 (Ralph, Michael & Han 1973, 9). Some Roman sherds found on the subsoil, described below, are probably to be associated with this hut. Phase 2: broch construction: Although the severe demolition of this fort at a later time means that the original wall survives for only part of its circumference, and to a height of less than 1.5 m, its general size and shape, together with the surviving architectural features (fig. 3), suggest that it was once a solidbased broch tower similar to the better preserved one at Torwood a few km to the east (RCAHMS 1963, no. 100). Also important for the inference that the Leckie structure was once towerlike are signs of crushing in the sandstone slabs at the base of the inner face of the wall on the N side that a great weight once rested on them (pi. 2) and the evidence for the demolition of a high wall which is outlined below. The internal court varies in diameter from 9.05 to 9.65 m and is slightly squarish; the markedly irregular shape of this broch is due to its cramped situation on the end of a narrow, steepsided promontory (fig. 2). The wall base is up to 5.8 m thick on the NW and 4.75 m on the NE. An intramural stairway oftypical broch design, and having a guard cell at its foot, is on the NW side (pi. 4), but no traces of an upper gallery were preserved on the wallhead, nor was there any sign of a scarcement on the inner wallface. The date of the building of this stronghold is indicating 61 Location maps m _» FlG.!. Location maps for Leckie site. ^

3 presumably fairly clear from the Roman finds. Fortunately two pieces of Samian ware were stratified unequivocally under the primary broch floor deposits and rested directly on the reddish sandy subsoil; both were of Flavian date (c. AD 80100) and one was a piece of a hemispherical bowl of south Gaulish design with the potter's stamp preserved. This has been read by Mr Hartley as probably FLAVIUS AUSfTINUS?) and the piece can be dated to the period AD Most of the pottery found in the primary broch levels is of the same date (Table 1) and the chronological significance of this is discussed below. It is clear that the Leckie broch was built after, and quite possibly soon after, AD 79 or 80; that would presumably be when the first Roman material became available to the natives of central Scotland as Governor Agricola's armies advanced. Phase 3: primary use of broch: The tower seems to have been inhabited for about halfa century, perhaps as much as 60 years (see below), during which time up to 30 cm of black earth accumulated on the original rock and soil surface, and over the strips of peripheral paving which had been laid on this in places around the wall. From the beginning the building had a large; central stone hearth (pi. 3), lying over one of the prebroch postholes, and it may therefore be classed as the dwelling of a single household. In this way Leckie contrasts sharply with the Dun Mor Vaul broch on Tiree which originally lacked this feature and seems to have been built to serve as a communal fort (MacKie 1974, 85) but resembles the semibroch (prototype broch) at Rhiroy in Wester Ross (MacKie 1981). A circle of eight massive postholes was found cut into the rock and subsoil, at a distance of from 1.8 to 2.3 m from the base ofthe wall, and they confirm that Leckie was indeed a broch. Apart from Iron Age wooden round houses (and the Rhiroy semibroch) only these tower forts were consistently equipped with an internal ring of posts, which probably supported a raised, annular floor (which also would have rested on the stone scarcement) and perhaps also a low roof. As the Table shows Roman material came from Phase 3 deposits but this is best considered together with that from the destruction layer which marks the end of that phase. End of Phase 3: destruction: On top of the primaiy broch floor there were clear signs of burning in the form of many fragments of fired clay daub from internal wattleanddaub screens and of a heap of charred barley near a rotary quern. These a C14 date of AD 45 ± 120 (GX grains gave 2779), equivalent to a calendar year time span centred on about AD 70 (Ralph, Michael & Han 1973, 9). Scatters of barley grains were found over large parts of the interior and neatly defined the destruction level in many places. The massive wooden posts of the main ring were pulled bodily out oftheir sockets at this time and several fresh pieces of fired daub and some entire artefacts fell into these together with black earth from the floor deposits. Immediately afterwards the high broch wall seems to have been thrown down to within a few courses of the floor and to have been totally obliterated along an arc from the NE to the SE; the main entrance was almost certainly here, facing a narrow path and a steep drop to the stream. There is a striking parallel with the discoveries made at the Torwoodlee broch in Selkirkshire (Piggott 1953), and probably also with those at the neighbouring Buchlyvie broch, 10 km W of Leckie (Main 1979). Several heavy stones which had evidently fallen from a height were found still lying at the base of the inner wallface on the NE, and others seem to have hit the hearth and damaged it (pi. 3). Traces of a mass of dry rubble were found over most of the central court, neatly stratified between the primary and the secondary floor deposits and on top of the fired daub and charred grain. There were many clues to the nature of this destruction which was clearly sudden; large numbers of complete ornaments, tools and weapons were lying on top of the broch floor, including bronze spiral fingerrings, ringheaded pins and penannular fibulae, an enamelled RomanoBritish bronze fibula, a fine bronze double chain brooch with enamelled ornament, two lead lamps and many lumps and splashes of melted lead, Roman and native glass beads, glass bangles and a Roman metal mirror. There were also many well preserved iron objects like a pair of sheep shears, a foot plough areas chrom, a sword blade, a cauldron hook and several small spear heads and knife blades (MacKie 1979, fig. 1). At least two curious rounded stream boulders 62

4

5 were found in the destruction levels different from the mass of tabular slabs of red sandstone which made up the wall debris of this quite phase and they had evidently been heated red hot and then suddenly cooled; they exhibit the characteristic heatcracking seen on pot boilers but on a much larger scale. One possible explanation is that they were red hot missiles hurled over the wall into the top of the broch by some form of artillery and doused with water by the defenders. Such missiles could have caused the lire. The latest dateable material from the broch levels is a piece of Samian pottery belonging to the period AD , and presumably available in central Scotland only after the Roman army reconquered the southern halfof the country in AD 140, in the reign of Antoninus Pius. This piece was slightly below the destruction level but three more Antonine fragments found were in it; however most of the Phase 3 Roman pottery was of Flavian (late first century) date. The shattered mirror is of first century manufacture and, like the Flavian pottery and glass, must have been kept safely inside the broch for many years. This broch tower, like that at Torwoodlee, seems to have been knocked down early in Antonine times, presumably soon after AD 140, at the time ofthe military reconquest ofsouthern Scotland. Phase 4: secondary occupation: Most of the fallen debris seems to have been cleared out of the wrecked broch some time later, and some of the stone was used to rebuild the wall in much cruder masonry, and to a height ofabout 3.0 m (pi. 2). A new oval fireplace was laid over the shattered broch hearth but for some reason the mass of tools and ornaments lying a few centimetres below the newly cleared surface was not recovered. This suggests that the new occupants, who had evidently converted the ruins to a roundhouse, were not connected with the original broch owners and did not realise what was still lying inside the ruins. A layer of lighter grey earth, mixed with some large stones, began to accumulate in the interior and a third, square hearth eventually replaced the oval one (fig. 6); it was itself contracted and modified right at the end of the site's history. The nature of the roof of the reconstructed dwelling is not clear; although it must surely have been of thatch on a wooden frame no traces of postholes were observed in the Phase 4 floor deposits. Dating evidence for Phase 4 consists of some Samian pottery of Antonine date (about AD ), two Flavian fragments and two silver denarii; these last were oftrajan (minted in about AD 105) and, surprisingly, of Julius Caesar and minted in about 45 BC. The coins, especially the Republican one, were obviously already very old by Phase 4 times although Trajanic coins are not uncommon on Antonine sites (Robertson 1963, Table). Phase 5: refortification and abandonment: There were clear signs that the roundhouse was partly refortified at the end of the period of secondary occupation. A start was made on converting the south side of the buildingthat facing along the flat and vulnerable approach along the promontory into a defensive crosswall; however the wallface was only three courses high in the middle and there was hardly any stone debris in front of it (pi. 1). Only the foundations for the heavy blocks had been laid at the west end. Large stone lintels were evidently being ripped out of the broch wall over the stairway and dragged round to the front; a few remained in position over the stair but had been dislodged at their outer ends and had tipped down into the passageway; there they rested on a thin layer of black soil, presumably secondary. Indeed the whole stairway seems to have been undermined by the removal of the outer wallface at this time, several of the steps being wrenched out of position and others cracked across by downward pressure on their outer ends (pi. 4). Some large animal bones from rubble associated with the dislodged stair lintels gave a C14 date of AD 110 ± 150 (GX 2780), equivalent to a calendar year time span centred on about AD 50. The interior wallface had apparently already been heightened with blocks of white sandstone, obtained from some distance away on the hills above, and these contrast sharply with the red stone with which the broch and roundhouse had been built. All this later stonework collapsed into the interior, falling directly on to the latest hearth and on to a number of artefacts lying on top of the secondary floor deposits. These last included a pair of iron hubrings from a cart or chariot wheel, the bone ballshaped head of an iron pin (of a type found in the brochs of the northern islands) and several pieces of bronze which may be vehicle fittings. 64

6 i i i «f> i i i i m i i i r LECKIE Phases 2 and 3 FlG. 3. Leckie: structural phases and external trenches. LECKIE Phase 1 Fig. 5. Leckie: plan of interior in Phases 2 and 3. LECKIE Phase 4 9 SUkBholes FIG. 4. Leckie: plan of interior in Phase 1. The site seems to have been completely abandoned after this final destruction or collapse, although a greenglazed sherd and a mediaeval longcross silver penny from near the surface of the deposits at the centre of the court show that it was visited some centuries later, presumably well before the recent rhododendron thickets completely covered the site. Discussion The southern brochs: The explanation of the scatter of brochs in central and southern Scotland (fig. 7) has often intrigued archaeologists (Childe 1935, 205; Piggott 1953, 11315; Stevenson 1966, 35). They lie far from the areas of main concentration of the tower forts in the Western Isles, in Sutherland and Caithness and in the Orkney and Shetland 0_ s_10 m Fig. 6. Leckie: plan of interior in Phase 4. and the discovery of quantities of late Islands first century Roman pottery and glass inside the Torwoodleee broch, Selkirk, 90 years ago (Curie 1892) showed that there must have been some link between the appearance of brochs in the southern mainland and the first penetration of this area by Roman forces in AD 79 or 80. Torwoodlee broch was reexcavated by Piggott in 1950 and proved to have been deliber 65

7 ately destroyed, the high wall being hurled into the surrounding ditch (Piggott 1953). The excavator surmised that the evident violence of this destruction meant that it was the work of the Roman army; a rival Iron Age chief would surely have taken the stronghold for himself. The evidence from Leckie strongly supports this view and we may even have from the Forth Valley site examples of the red hot stone shot hurled into the tower by a piece of artillery to set it on fire. If so one suspects that when flames and smoke filled the interior surrender quickly followed. The date of the construction of Torwoodlee broch seemed clear from Piggott's discovery that Flavian Roman pottery and glass had been on the site before the tower was built; pieces were found under the wall. Piggott followed Curie in supposing that this broch, and therefore presumably the whole south Scottish group, had been built after the first Roman withdrawal to the TyneSolway line in about AD 100; both excavators agreed that the power vacuum in the territory of the Votadini (and elsewhere in the south) which was presumably thus created would have been a suitable moment for adventurous broch chieftains from the north to seize some new territory. In this scenario they would have obtained their 66

8 presumably Roman material from the now abandoned fort at Newstead, a few kilometres to the east. The destruction of Torwoodlee, and presumably of other southern brochs, would then have occurred less than 40 years later when the Roman army was preparing for its second, Antonine occupation of southern Scotland. The chronology of Leckie: The more precise dating for the Leckie broch allows this picture to be amplified a little. That the destruction did indeed take place soon after the start of the Antonine period is clear from the latest pottery in the Phase 3A levels. Leckie, like others in the Forth Valley such as Goldoch perhaps (Graham 1949, 12) and probably Buchlyvie (Main 1979), was doubtless destroyed specifically to protect the network of Roman roads and bases being reestablished north of the Antonine wall construction at that time under and to which end control of the was essential. Stirling crossing This chronology fits well with subsequent events. The secondary occupation in Phase 4 was clearly contemporary with the Antonine period of about AD (Table 1) and the final abortive refortification could have been a response to the campaigns of Septimius Severus and Caracalla in AD (fig. 8). As far as the date and circumstances of the construction of Leckie are concerned however, the new evidence suggests a new conclusion. This is that some at least of the south Scottish brochs were set up by warrior chieftains and their armed followers and households who had come down from the far north and who had established themselves, with the agreement of the Roman authorities, some time during the Flavian period. In fact a suitable time for this to have happened would have been in the mid 80s, after Agricola's recall, when the Romans abandoned their forts and roads north of the ForthClyde line. If this was so the broch people would have been at first allies of Rome, encouraged to come south to form in effect small buffer states immediately north of the frontier and perhaps, in the case of the Phase Events & Features C14 dates Closely Daleable Material Roman Samian ware Other Imported Material Roman glass Lead Xutiv Iron To oo n O cn 2 % End of Secondary Occupation (abortive refortification) AD 110 ± 150 (GX2780) Secondary Occupation Destruction (end of primary use) Primary Occupation 1 Hadrian (c. 105) 1 Caesar (c. 45 BC) AD 45 ± 120 (GX2779) I Construction cs O Wooden Hut AD 80 ± 70 (GU1370) Table 1: The Leckie broch: important finds related to stratigraphy. 67

9 Selkirkshire brochs, to help hold down potentially rebellious southern tribesmen. There is much circumstantial evidence supporting this view. Essentially it depends on whether Flavian Roman material in the quantities found at Leckie and Torwoodlee is likely to have been obtained by trade during the late first century occupation Roman or as loot a short time later from abandoned Roman camps. The former view now seems much more plausible. The Romans tended to bury material they could not take with them on their withdrawals, like the well known nails at Inchtuthil, whole Samian pots in the ditch at Cardean (Robertson 1978) and probably all the stone monuments which once embellished the Antonine Wall (Keppie 1979). It has been demonstrated that the range of Roman coins found on native sites is almost identical to those from Roman ones, showing clearly that currency flowed directly from Roman to native hands (Robertson 1963, Table). The same would surely apply to pottery, glass vessels, beads and so on the vast majority of which should only have been transferred when Romans were on hand to provide it. The complete absence of Hadrianic Roman material of the period AD from all the occupation layers at Leckie which must have accumulated continuously apart from a short interval after the destruction of the broch is very important and surely confirms that the occupants obtained their imported material from traders only while the Roman army was in the neighbourhood. I am indebted to Mr Peter Webster for this vital deduction. The numismatic evidence does not show this so clearly (Robertson 1963, Table), early second century coins of Trajan and Hadrian being common on Antonine Roman and native Iron Age sites alike. However old coins obviously circulated for many years in Roman pockets whereas Samian pots and glass vessels, we may assume, were subject to rapid breakage; they surely reflect as whole imports more accurately the date of the trading contacts. There must have been many merchants and traders moving among the tribes immediately beyond the Imperial frontiers and powerful native chiefs would have been obvious customers. Doubtless the arrival of Agricola's army in AD 79 or 80 encouraged traders to search for custom among the chieftains of the Forth Valley and beyond, especially if as is suggested here the brochs were newly built and held by northern people friendly to the invader. We know from Ptolemy's Geography ofbritain compiled early in the second century (Rivet 1978) that the names and positions of many settlements and natural features all over highland Scotland and Ireland were known to the Romans and, though much of this knowledge must have come from the army and the fleet, much also and presumably all ofit from Hibernia and explorers. must have come from traders Allies of Rome? That the southern brochs were built according to the traditions of the far north rather than ofthe Hebrides is clear from their architecture; they all seem to be of the solidbased form whereas Hebridean brochs, with two exceptions (Dun Troddan and Dun Telve), are of the groundgalleried type (MacKie 1965, 107). The best preserved of the southern brochs is Torwood, about 15 km SE of Stirling (RCAHMS 1963, no. 101) and it has a combination of structural features the ledge type scarcement and the massive solid wall base which is found elsewhere only in Sutherland, with a few in Caithness and Orkney (MacKie 1974a, figs. 6 & 15). Some of the native artefacts at Leckie are of northern type too, particularly the bone head from an iron pin already referred to. V G Childe referred many years ago to the statement of Orosius that the chieftains of Orkney had made formal submission to the Emperor Claudius at the time of the invasion of England in AD 43 (Childe 1946,129); he thought that these chiefs, who must have been broch owners at that time, had therefore but recently arrived from the S since they were familiar with a fairly sophisticated diplomatic manoeuvre practised by Gaulish tribes about to be overrun by the advancing Roman army and who did not wish to be devastated in the process. This interpretation of Orosius has been doubted (Maxwell 1975) but we also know that Agricola's fleet circumnavigated Scotland during his 5th campaign in about AD 83 or 84; the ships must then have made contact with the brochowning tribes of the far north. It may not be too fanciful to suggest that the Roman authorities persuaded some of these already perhaps with a history of diplomatic contact with the Empire to come south and, like the AngloNorman 68

10 knights of a millennium later, to help keep recently subjugated tribes quiet. If so this policy had to be changed at the time of the Antonine invasion and the brochs were destroyed. The map shows how, of the four separate groups of brochs in the centre and south of the Scottish mainland, the two most northerly lie immediately beyond the main Roman frontier from the Forth to Clyde in and near the Forth Valley, and around the Firth oftay respectively. Their situation fits very well with the hypothesis offered here. Whether the small group in SE Scotland can be explained in the same way is not so clear; it might be objected that the building of native towerforts behind the frontier instead of immediately in front of it, even by allies, is unlikely to have been allowed so readily by the Roman authorities, yet the quantities of Roman material found inside Torwoodlee give impression a similar as Leckie that of a chieftain enjoying favoured status and many luxury imports. FlG. 7. Map of brochs in southern Scotland. 69 Perhaps these southeastern brochs were built in about AD 100, again with the agreement of the Roman army, by groups from the Forth Valley and Fife strongholds just as the military was withdrawing from Scotland. As far as the two or three possible brochs in the extreme southwest of Scotland are concerned, neither has been excavated so nothing definite can be said about their date. Roman military sites certainly existed not far away to the east. Implications forfinds: The sudden destruction of the broch at Leckie seems to have resulted in the preservation of a large part of the imperishable equipment of one wealthy Iron Age household of the mid second century, and much useful information can be extracted from it. For example we now know, contrary to some previous impressions, that these people were fully armed with iron swords and spears; they also had a fine array of native

11 4 ABANDONED contac r? (severan) t no contact ROUNDHOUSE contact wi th rome (anton/ne) 3a BR0CH N no contact USE BRQCH BUILT.HUT. conta ct with rome (Flavian) pre roman FlG. 8. Table of structual phases and Roman contact (the dark strip is phase 3b). 70

12 jewellery and imported luxuries and seem to have lived well, either by growing their own barley and tending flocks of cattle and sheep or by taking these supplies as tribute from the local community. The existence of sheep shears and cas chrom suggests the former, that the broch household included active farmers and did not simply live off the tribute of subordinate tribespeople nearby. The occupants of this particular stronghold were wealthy enough to obtain quantities of exotic material, including a rare first century Roman mirror. It is interesting to note the usually high proportion of Samian to coarse ware on the site 8020% (58 of 73 sherds); this is much higher than would be expected in S Britain and again may reflect the high social status of the inhabi or military importance tants during the later Flavian period. Hardly any native sherds were found but wooden bowls were in use, the charred fragments of one appearing in the broch destruction level; thus it was probably not a question of Roman vessels being used instead of everyday ones but rather of such vessels being obtained as luxuries. If they were once complete the glass vessels imply that even more clearly and the remarkable array of native jewellery and Roman beads also indicates considerable wealth and prestige. Only the much larger Traprain Law hillfort in E Lothian has produced a richer haul of native and imported valuables of this period (Burley 1956). A full study of the Leckie finds will provide more information in due course but certain preliminary inductions can be made, based on the very clear and well dated phases of occupation (fig. 8); these tie down the chronology of some of the native artefacts very closely. In summary Phase 2 probably occurred during the AD 80s, Phase 3 lasted from then until about 140, and Phase 4 probably from the early 140s to perhaps about 205. The relative frequency of Samian sherds and of everyday native iron artefacts (Table 1) shows that we would expect to find about three times as many of any particular artefact in the broch levels (Phases 2 and 3) as in the postbroch deposits (Phases 4 and 5) if they were equally used in both periods. However certain objects are far more frequent in the broch period and their abundance tells us more about the nature of that occupation. Glass armlets: These are thought to have been made of Roman glass with native enamel decoration (Stevenson 1974) and are almost entirely confined to the broch period. Most of the fragments with ends preserved, even the one from Phase 1 or 2, show collars as if broken armlets had been reworked to be set in a mounting (Stevenson 1974, 523). It may be suggested now that the Scottish Iron Age glass armlets the final northern flowering of a Continental La Tene fashion were probably made during the short period of the Flavian Roman occupation at the end of the first century, when quantities of the raw material suddenly became available, and that many were quickly broken and reused. They seem to have almost entirely gone out ofuse by Antonine times. The whole phenomenon, one suspects, is another clear sign of the favoured treatment of some Iron Age tribes by the Romans, and also perhaps of the close connections that the north British tribes once had with western Gaul. Lead: An extraordinary quantity of lead was found at Leckie, mostly in the form of lumps and ingots but including some artefacts like whorls, weights, and two lamps which are metal versions of the Iron Age stone lamp; there was even a small piece of galena (lead ore). The only source for so much lead must have been Roman traders, or perhaps the military authorities since it was imported for use as drains and other things in Roman arjny forts and bathhouses (Keppie 1981, fig. 20). This is perhaps one of the clearest signs of close and friendly links between the broch owner at Leckie and the Roman military authorities of the Flavian period. Hardly any lead came from the postbroch levels, a vivid indication of the change in status of the community in the eyes of Rome. Iron artefacts: The quantity and quality of the iron tools and weapons preserved at Leckie is unique for a native site in Scotland; it must have something to do with the welldrained nature of a site built on a sandstone promontory. Some of these have already been illustrated and briefly described (MacKie 1979, fig. 1) and the final report will contain full details. Leckie is plainly an important site in terms of the known historical events of Scotland in the first and second centuries AD, and it will also provide a large amount of information about the technology, economy and living standards ofone wealthy Iron Age community of the period. 71

13 Acknowledgments I am very grateful to Mr Peter Webster and Mr Brian Hartley for their reports on the Roman pottery, to Dr Donal Bateson for his opinion of the coins, to Miss Glynis Lloyd Morgan for studying the mirror fragments and to Dr Lawrence Keppie for much help with all the Roman material. The conclusions drawn here about the place ofthe Leckie broch in the history of Roman Scotland would not have been possible without the detailed dating of the Roman finds provided by these scholars. Bibliography Breeze, D J (ed) 1979 Roman Scotland: some recent excavations. Newcastle. BURLEY, E 1956 'A catalogue and survey of the metalwork from Traprain Law', Proc Soc Antiq Scot 118 (195556), Childe, V G 1935 Scotland in Prehistory. London. Childe, V G 1946 Scotland before the Scots. London. Curle, J 1892 'Notes on two brochs recently discovered at Bow, Midlothian, and Torwoodlee, Selkirkshire'. Proc Soc Antiq Scot 26 (189192), graham, A 1947 'Some observations on the brochs', Proc Soc Antiq Scot 81 (194647), Graham, A 1949 'Notes on some brochs and forts visited in 1949', Proc Soc Antiq Scot 83 (194849), Keppie, L J F 1979 Roman distance slabsfrom the Antonine Wall: a brief guide. University of Glasgow. Keppie, L J F 1981 'Excavation of a Roman bathhouse at Bothwellhaugh, ', Glasgow Arch Journ 8, MACKlE, E W 1965 'The origin and development ofthe broch and wheelhousebuilding cultures of the Scottish Iron Age', Proc Prehisl Soc 31, MacKie, E W 1974 Dun Mor Vault an Iron Age broch on Tiree. University of Glasgow. MacKie, E W 1974a The origin and development of the broch and wheelhousebuilding cultures of the Scottish Iron Age. PhD Thesis: Glasgow. MacKie, E W 1979 'The origin ofironworking in Scotland', , in M Ryan (ed) Proceedings of the Fifth Atlantic Colloquium. Dublin. MacKie, E W 1981 'Dun an Ruigh Ruaidh, Loch Broom, Ross & Cromarty: excavations in 1968 and 1978', Glasgow Arch Journ 7, Main, L 1979 'Excavations at the Fairy Knowe, Buchlyvie, Stirlingshire', in D Breeze (ed) 1979, Maxwell, G 1975 'Casus Belli: native pressures and Roman policy', Scott Arch Forum 7, PlGGOTT, S 1953 'Excavations on the broch and hillfort of Torwoodlee, Selkirkshire, 1950', Proc Soc Antiq Scot 85 (195051), RlVET, A L F 1978 'Ptolemy's Geography and the Flavian invasion of Scotland', Studien zuden Militargrenzen Roms 2, Robertson, Anne S 1963 'Roman coins found in Scotland, ', Proc Soc Antiq Scot 94 (196061), Robertson, Anne S 1978 'The Roman Fort at Cardean', in DJ Breeze (ed) 1979, 424. RCAHMS 1963 Stirlingshire: an inventory of the ancient monuments. (Royal Commission on the Ancient and Historical Monuments of Scotland). HMSO: London. Stevenson, R B K 1966 'Metalwork and some other objects in Scotland and their cultural affinities', in A L F Rivet (ed) The Iron Age in Northern Britain. Edinburgh. Stevenson, R B K 1974 'RomanoBritish glass bangles', Glasgow Arch Journ 4, This paper is published with the aid of a grantfrom the Carnegie Trust for the Universities of Scotland. 72