Irish souterrains: Later Iron Age refugees

20
R Warner Irish Souterrains: Later Iron Age refuges. Archaeologia Atlantica 3,1980, 81-99. For Later Iron Age read Early Medieval (the presently accepted term for the period) throughout.

Transcript of Irish souterrains: Later Iron Age refugees

R Warner

Irish Souterrains: Later Iron Age refuges.

Archaeologia Atlantica 3,1980, 81-99.

For Later Iron Age read Early Medieval (the presently accepted term for the period) throughout.

IRISH SOUTERRAINS: LATER IRON AGE REFUGES Richard Warner

Among the rich surviving monumental cultural heritage of Ireland one of the best preserved, commonest and least studied monument types is the artificial cave usually called, in recent lite­rature, by the borrowed French term 'souterrain'. This important antiquity has received scant attention in most archaeological texts, usually warranting only a few lines of not very useful comment. There have been only two reasonably thorough general discussions: by S.P. 0 Riordain* and by the present writer ^.

DESCRIPTION An Irish souterrain can be described as an artificial subterranean structure built to allow restricted access, and temporary accommodation for a number of people. Four types of unit, or module, are found to be sufficient to describe most souterrains: the passage, the chamber, the creep (a restric­tion or impediment) and the entrance. In their basic provision of underground shelter (argued be­low) all souterrains are very much the same, and indeed there is remarkable similarity among the simple examples because of the 'modular' nature of their design. Complexity and variation were obtained by the number of times the modules were used in a particular souterrain, and the rela­tionship of the modules to one another. Souterrains may be primarily classified according to their basic constructional technique, although examples may be found in which two or more techniques are combined. Sub-classification is then possible by analysing the way the basic units have been used, but I have not attempted to do this here. 1. Tunnelled Souterrains These are found wherever the bed-rock or sub-soil is firm enough for tunnelling. The simplest are straightforward 'burrows' whose spoil was removed through the mouth, and are necessarily limited in their complexity (fig. 5 upper). Nevertheless the longest Irish souterrain known is of this type, being at least 130 metres long with at least eight oval chambers, each separated from the next by a constriction of a type that I shall refer to as a 'simple creep'. This particular souterrain runs under the Ringfort (see below) of Rathmore, Co. Antrim, historically attested as the capital of a powerful tribe (Dal n'Araide) during the second half of the first millennium A.D.^. In the more sophisticated 'spoil-pit'^ class of tunnelled souterrain (fig. 1) the passages and chambers were dug from the base of a wide, deep pit, the spoil being removed through the pit. After completion of the tunnels the openings at the bottom of the spoil-pit were blocked off with dry-stone walling (sometimes incorporating an air-shaft) and the pit refilled to the surface with the spoil. Short souterrains, or those with a circular overall plan, were often dug from a single spoil-pit, but more extensive structures often required a number of such pits. The entrance to a tunnelled souterrain

82 Richard Warner

can be steep and narrow, often long and often sealed at the mouth by one or two stone slabs'. Tunnelled souterrains have been found to a depth of 3 metres or more below the surface. Features can include stone-buUt and rock-drilled air-shafts, raised platforms and 'child holes' (?). Fine examples with these features have been described at Curraghcrowly, Co. Cork, and Ballintemple, Co.Derry^ 2. Dry-stone Souterrains^ Soutenains of this constructional class are usually found where the natural rock or clay is unable to sustain a tunnel, but because the technique allows very sophisticated and complex variations they are also found where tunnelling would have been possible. A dry-stone souterrain was con­structed, in a square-sectioned trench, of dry-stone walling usually slightly corbelled inwards, capped by large stone lintels and topped with soil to the surface, which is seldom more than 1 metre above the bottom of the Untels. In this class two types of chamber are encountered. The long-oval, or sub-rectangular, chamber is typically 1.5 to 2 metres high, 1 to 2 metres wide and up to 3 metres long (fig. 2 upper). It can be entered through the end or through a long side, or occa­sionally through the roof or floor. The roof is simply lintelled, with a minimum of corbelling. The circular 'beehive' type of chamber, usually about 2 metres high and wide, has a corbelled roof, topped by one or two lintels, giving a hemispherical shape to the chamber (fig. 2 lower)*. The two chamber types, beehive or oval, are seldom found together in the same souterrain. Souter­rain passages vary from the 'restricted' type, hardly one metre square yet up to 20 metres long, to the 'roomy' type, up to 2 metres high and 1.5 metres wide (as large as chambers) and can be straight, angled or curving. Creeps, which are structures built to impede movement within the souterrain, are typically found between chambers, between passage and chamber, subdividing a passage or at the entrance. They are usually of the 'simple' variety, a short, lintelled tunnel less than one metre square. The more complex 'drop-hole' creep is a small square opening in the floor of one chamber or passage and the roof of another beneath. A person might rise or fall through a drop-hole creep, depending usually on the slope of the ground in which the souterrain was built. A comphcation of the drop-hole creep is the 'bridge-creep'' in which a rising drop-hole creep enters a short stretch of passage, followed shortly by a falUng drop-hole creep (fig. 3). A number of other features can be found in dry-stone souterrains. 'Air-shafts' (or air-vents) are narrow stone-built pipes running upwards from near the roof of a chamber or passage to exit at some (often hidden) point above ground*". Internal wooden doors, particularly next to simple creeps, have been indicated by pairs of post-holes". 'Escape tunnels' have been f o u n d l e a d i n g from an inner, or end chamber to the outside (for instance, to the face of a nearby steep slope, or the ditch of the fort). Less common features include drains (narrow stone-buUt pipes leading away from the base of a chamber wall, downwards), platforms, benches, shelves, alcoves and recqsses. The entrances to stone-built souterrains are occasionally simple sloping ramps or steps leading to the first creep, or even straight to a passage or chamber. More commonly, and particularly evident when proper excavation has been undertaken, the entrance is found to be a steep or vertical shaft leading to a creep, or sometimes simply a hole in the roof of the first chamber or passage.

Irish Souterrains: Later Iron Age Refuges 83

B A L L I N T E M P L E

d r y - s t o n e blocking

h chi ld h o l e ?

C U R R A G H C R O W L Y RBW

Fig. 1. Spoil-pit tunnelled souterrains at Ballintemple, Co. Derry (after May and Cooper**), and Curraghcrowly, Co. Cork (after Somerville*).

3. Wooden Souterrains A number of souterrains have been excavated in which the presence of post-holes, the absence of lintels, and vertical or absent side walling, in an otherwise normal souterrain trench must be inter­preted as indicating the former existence of a trench-built wooden-roofed, or completely wooden

84 Richard Warner

souterrain. In general the wooden souterrains seem to have been very similar to the dry-stone ones, though perhaps simpler. Several were found, for instance, in the ringfort of Ballycatteen

CHRONOLOGY The vast majority of souterrains have produced no dating evidence of any kind, and records of finds of artifacts have often been inadequately recorded. I have explained elsewhere why we need take no notice of the single piece of apparent evidence for a Neolithic date*'*, but two souterrains have been held to date to the Bronze Age. One, an undescribed 'cave' in Co. Co^k*^ is reported to have produced 'from a shelf a Middle Bronze Age palstave and a Late Bronze Age socketed axe. There is no reason why the two objects should not have been associated with each other ** but there is no evidence that the cave was a souterrain. In fact the immediate area is rich in early copper mines and copper ore sources*^ and a copper mine would seem a much more likely ex­planation of the 'cave'. The other site was Cush, Co. Limerick, an interconnected group of ring-forts**, one of which (no. 5) was found to contain a number of burials dated by the excavator to the Late Bronze Age*', but which would now be called Early Bronze Age °. One of these burials was claimed to post-date, stratigraphically, a souterrain and round-house and thus (accor­ding to the excavator) the ringfort. The finds from the ringfort would not now be dated by any competent archaeologist to earlier than the beginning of our era. Most would date them, and thus the whole group of ringforts and souterrains (one of which was shown to be stratigraphically earlier than no. 5), after A.D. 500. Cush therefore is unacceptable evidence for the Bronze Age or Early Iron Age date of a souterrain. Where souterrains have been found to have a true contemporaneity with any settlement, and a large number have now been adequately excavated as part of the excavation of larger habitation sites, the settlement has proved to be either directly datable to the Later Iron Age (c. A.D. 500 to A.D. 1200), or of a type which would belong only to that period. Despite suggestions to the con­trary, to be found in most Irish writings on these settlement types^*, no true ringforts or souter­rains have yet been satisfactorily shown to pre-date the 5th century A.D., or to have been con­structed after the 12th^^. Individual finds from souterrains support this view; a circa 9th century 'Teutonic' beaker from Mixllaghroe, Co. Sligo^^, 10th century Saxon coins from souterrains at Knovifth, Co. Meath** and Dundalk, Co. L o u t h 6 t h or 7th century radiocarbon dates from a wooden souterrain at Raheennamadra, Co. Limerick^*, a Later Iron Age pin sealed by the con­struction debris of one at Letterkeen, Co. Mayo^'', Later Iron Age pottery associated with nume­rous souterrains in the north-east^* and inscribed memorial stones (both ogham-inscribed and cross-inscribed) of the same period from numerous s o u t e r r a i n s T h e Ust could be far longer, but these examples suffice to give the general picture. The other Une of evidence for the date of souterrains is the rich store of early Uterature, anna-listic and narrative, written down from the 7th century A.D., but reliably recording events back to the 6th ' ' ' . In this literature the souterrain figures prominently as a contemporary structure from, at the latest, the 9th century A.D., until the 12th'*. The Old-Irish word used is Uam, which means simply 'cave' but in almost all Uterary cases should be understood as 'souterrain'. I shall

Irish Souterrains: Later Iron Age Refuges 85

K N O C K D H U

O F = r 310

d r y - s t o n e wal l ing

clay and infill

air shaft

simple creep

drop hole creep

ent rance

•••e?

T E R M O N F E C K I N RBW

Fig. 2. Oval-chamber dry-stone souterrain at Knockdhu, Co. Antrim (after Lawlor'°). Beehive-chamber souterrain at Termonfeckin, Co. Louth (after 6 Floinn*).

86 Richard Warner

allude to some individual literary references in the foOowing sections. Finally, we have a very strong piece of evidence that the Irish souterrain does not predate the middle of the first millen­nium A.D. In the 4th and 5th centuries A.D. the Irish are known to have migrated from areas in which souterrains are now common (such as Co. Waterford and north Co. Antrim) to Bri­tain'^. But in the areas of their consequent settlement, such as east Cornwall, south Wales and south-west Scotland, souterrains are completely unknown.

DISTRIBUTION Souterrains have been found in almost every part of Ireland, over a thousand are known, or have been recorded, but their distribution is very uneven' ' . This is surprising, forjudging by the distri­bution of the contemporary ringforts the density of human habitation was fairly even, avoiding only land above about 200 metres in height and peat bogs'*. In Co. Antrim over 600 souterrains have been recorded, while in neighbouring Co. Tyrone less than ten are known. Even on a more local scale this marked unevermess is indicated'^. It is quite impossible to explain this phenome­non by topographical or geomorphological means, as can be done with the uneven distribution of the contemporary lake-dwellings. Nor can an unevenness of research or farming practice be called upon to explain it, there being no significant correlation between intensity of fieldwork or land disturbance and areas of souterrain clustering. It is possible that a purely morphological explanation may account for the irregularity, if those areas lacking the stone-built or tunnelled souterrains (which are those easiest to find today) had, instead, larger numbers of wooden souter­rains, which would not now readily appear except during excavation. As a number of contempo­rary sites have been excavated in such areas and have failed to produce signs of even a wooden souterrain, and the wooden souterrains that have been found have been in areas already known to contain stone-built or tunnelled ones, this seems an unlikely explanation. I will suggest below that the souterrain is a simple response to a particular socio-economic need, but it is not possible to call upon a socio-economic explanation for the high degree of unevenness in the Irish distribution. The same level of wealth, and the same cattle-based economy seem to be apparent in the areas in which souterrains are rare as in those in which they are plentiful. It is tempting to suggest that there might be a cultural explanation, but it proves extremely hard to find any other aspect of Later Iron Age material culture which takes any cognisance of the tribal and ethnic boundaries in Ireland. For instance in the north-east the contemporary 'Souterrain-ware' pottery is distributed quite evenly over the whole region (except the high land), showing no major local variations and occurring without distinction on sites with and without souterrains'*. In addition some souterrain clusters cross important tribal, ethnic and political boundaries. Never­theless it must remain a possibility that its failure to spread generally was due to a strong cultural or religious antipathy, amongst some peoples, to going underground. I will pursue this possibi­lity later. Although all the features described can be found in most of the areas in which souterrains occur in Ireland, certain repetitive associations occur, and in some cases the 'types' thus identified appear to be localised. The identification of these regional types and groups is only recently under

Irish Souterrains: Later Iron Age Refuges 87

Fig. 3. Dry-stone 'roomy passage' souterrain with bridge creep at Donaghmore, Co. Louth (after Rynne ' ) .

88 Richard Warner

way but one or two can be mentioned. The 'spoil-pit' tunnelled souterrains are commonest in Co. Cork, and are extremely rare outside it. In Co. Meath on the other hand (for instance at Knowth) the usual form is a long restricted dry-stone passage with one or more beehive chambers. In Co. Louth and the adjoining part of south Co. Down the long roomy passage, with drop-hole or bridge creeps and often no chamber (simply a widened passage end) is particularly common. In Co. Antrim and east Co. Derry the usual souterrain has an oval chamber, or several separated by simple creeps. There is no evidence, such as coinciding with known territorial extents, that these morphologically distinct groups are culturally or tribally determined. It is probable that the explanation of these differences lies in the professional status of the souter­rain builders (payment for a souterrain was two cows, the same as for a stone ringfort or a cause­way'"'). Each morphologically and geographically distinct group would show the area of influence or activity of a particular school or master. It is interesting that the Co. Meath type bears a sort of resemblance to the passage and chamber of the Late Neolithic Boyne-valley passage-graves'®, and at two of these (Knowth and Dowth) the passage and chamber are incorporated into the souterrain complex of their associated, though of course much later, settlements. It has been suggested that this shows the survival of a megalithic tradition ('submegalithic')" over the intervening three or four thousand years, an idea with absolutely nothing to commend it. Much more likely is that a local group of builders, having no tradition, copied the only local model of dry-stone building they could find. It might be thought reasonable to suppose that the gaps in the distribution were a reflection of the extent to which the activity of the 'schools' had spread by the time of the 12th century Anglo-Norman invasion, bringing quite different social and economic problems to the country. For such an explanation however we would have to postulate an almost explosive in­crease in the popularity of the souterrain shortly before this event, creating the extreme density observed in certain restricted areas, a situation which seems unlikely.

CONTEXT The majority of souterrains, found by chance (such as the collapse of the roof), are without any obvious archaeological context, such as traces of habitation. Nevertheless, where excavation has taken place on such souterrains they have invariably been shown to be associated with habitation, either by the finding of the remains of a building, or from scraps of domestic refuse or from the recognition of enclosing walls or banks. It is therefore reasonable to claim that all souterrains were adjuncts to habitation sites, and the finding of an apparently unassociated example may be taken to indicate the presence of an otherwise unknown settlement. A particularly important and widespread settlement type of the Irish Later Iron Age is the ringfort, a defensive enclosure of circular shape, surrounded by one or more concentric ramparts. These ramparts may be of earth, with an external ditch, or of stone, usually without a ditch. Excavation has shown that ringforts were defended farms, probably of free-farmers and the upper ranks of tribal society, the enclosure containing the main house and a number of subsidiary buildings'"'. Souterrains are commonly associated with these forts, usually being found within the enclosure. At Ballycatteen, Co. Cork'", the quarter of the multivallate ringfort excavated produced three souterrains, all apparently en­tered from the same large rectangular wooden house. At Cush, Co. Limerick (fig. 4)"*^, the indivi-

Irish Souterrains: Later Iron Age Refuges 89

C U S H _ _

/

B A L L Y W E E

0 10 20 30 1 I I I m .

ditch

mound

a i r - s h a f t

souterrain

building

^ entrance to souterrain

RBW

Fig. 4. Ringfort complex and souterrains at Cush, Co. Limerick (southern group; after () Rior­dain**). Dry-stone soutenains in mounds, with houses, at Balljnvee, Co. Antrim (after Lynn'°).

dual ringforts each contained at least one souterrain, other souterrains being found in undefended corners of the associated field system, though still adjoining buildings. It must be said that souter­rains (and ringforts) were usually placed in precisely the sort of topographical situation in which

90 Richard Warner

one would expect any other rural, pastoral-based, habitation. Relatively dry slightly sloping land, not too far from water, sited for agricultural rather than strategic convenience. Not all settlements were ringforts, or even single farm-S. There is growing evidence that undefen­ded settlements, whose discovery is often the result of the accidental finding of a souterrain, were extremely common, and, judging by the occasional report or discovery of several souterrains in close proximity to each other, these settlements might have been quite extensive, justifiably to be interpreted as villages. The large size of many Irish monasteries, or at least of their dependent settlements, is clearly shown in those cases where surrounding defensive earthworks survive, often enclosing several hectares*' and which, in some cases, we may justifiably call towns**. Souterrains are commonly associated with these monastic settlements*'. The association of the souterrain with a house is well demonstrated by excavation** (fig. 5). In the few cases where the shape and extent of the house were well preserved the souterrain can often be shown to have been entered from inside the house, but ran almost completely outside it. Occa­sionally the souterrain was actually incorporated in a thickening of the house wall*''. It seems Uke-ly that in those cases (the majority) in which all traces of the house have disappeared, the position of the entrance of the souterrain would be a good indication of the original position of the house. Occasionally a souterrain-like chamber or passage had been constructed within the rampart or wall of a ringfort or monastic enclosure. At Leacanabuaile, Co. Kerry**, the chamber was quite small, but was entered from below (through a rising 'drop-hole' creep) from a normal underground souterrain which ran from a circular stone house within the enclosure. Within the same rampart was another small chamber entered simply through a hole in the inner face of the rampart. This may well indicate the original existence of a peripherally placed wooden building. In a number of other sites*' the rampart chamber, or passage, in effect an above-ground souterrain, is found to run for a considerable distance around the enclosure, even completely around, and again peri­pheral buildings may well be indicated. There are, as we have just seen, a number of cases where a souterrain-like structure is wholly, or partly above ground, either covered by a mound of soil, incorporated in the wall of a house or built within the rampart enclosing the settlement (fig. 4 lower)'". It is possible that these were, in many cases, intentional devices (the opportunistic use of a rampart, wall or mound), but others may well have been forced upon the builders by a high water table, or bed-rock which was too hard to cut. There are also ringforts in which the interior was heightened artificially, either by a normal accretion of occupation, as at Cahercommaun'*, or intentionally to produce a 'raised ringfort', of which a large number are k n o w n S o u t e r r a i n s inserted into one of these, or into a pre-existing prehistoric mound such as JCnowth", I would regard as properly subterranean.

PURPOSE 1. Storage The most widely held view of the purpose of Irish souterrains, that they were primarily intended for storage with a possible secondary status as places of refuge, is one which I believe cannot be

Irish Souterrains: Later Iron Age Refuges 91

O L D C O U R T

Fig. 5. Tunnelled souterrain with round wooden house, in ringfort, at Oldcourt, Co. Cork (after Murphy and 0 Cuileanain''). Dry-stone souterrain with rectangular house at Craig Hill,

Co. Antrim (after Waterman'**). held when the evidence is properly studied. By far the majority of Irish souterrains are, or were when built, difficult to get into, often with an impediment placed just inside the entrance. Move­ment within was further hampered either by the use of a long restricted passage, or where the passage was of the 'roomy' kind, by the use of creeps. Even in those cases where the first chamber, or the part of the passage nearest the entrance, was relatively accessible it is unusual to find that the rest of the structure was equally so. It cannot be denied that small containers could well have been placed just inside the entrances of many souterrains, but the extreme difficulty of nego­tiating the majority of the known examples even without being hampered by baskets or pots

92 Richard Warner

makes nonsense of such a suggestion for the souterrain as a whole. It is obvious that a single-chamber cellar, with a strong door, is aU that is required for the storage of foodstuffs. Fragments of pottery and other non-valuable domestic objects have been reported from northern Irish souter­rains, but in almost all cases in a situation where they would best be interpreted as having fallen in either during or after occupation of the accompanying settlement. A few cases of whole pots or wooden vessels being found in the chambers, as at Shanneen Park, Co. Antrim'*, might best be interpreted as representing provisioning in times of danger. By far the majority of souterrains, even those whose wet conditions would have allowed the preservation of organic containers, have produced no finds whatsoever, and other signs of everyday use, such as fine ash or charcoal carried or blown in, are rare. Where objects have been found well within a souterrain it is usually clear from their nature and position that they were personal valuables hidden for safekeeping, probably at a time of danger. I have already mentioned the Saxon coins and glass beaker and add some more examples. A sou-terrain at Oldcourt, Co. Cork" contained, carefully hidden under a flagstone in one of the cham­bers, a bronze-coated iron bell. A souterrain in a stone-walled ringfort at Cahercommaun, Co. Clare'*, contained, carefully hidden near the end of a chamber, a 9th century A.D. silver brooch and a human skull with an iron hook (severed head and its means of suspension). The hiding of such objects would clearly be a sensible use of a souterrain, but not a reason for building such a large and complex structure. Any of these objects could as well have been hidden in a small hole in a wall of the fort, or house, as at Carraig Aille, Co. Limerick'''. No contemporary narratives, as far as I know, refer to the normal storage of possessions in souterrains, but several refer to the hiding of possessions in times of danger. For instance we read in one early story how the defenders of a ringfort collected the wealth of the fort together when they came under attack and deposited it in the souterrain'*. Another early source tells how, when famine hit Ireland (a frequent occur­rence) the people made 'strong deep cellars to save and hoard their victuals'". It is this danger-hiding of valuables that would have prompted the frequent plundering of souterrains, again at­tested in the early sources. We learn, for instance, that many souterrains in Co. Meath were plun­dered jointly by Scandinavians and Irish in A.D. 862, and those of Co. Kerry by Scandinavians in 866*". The souterrains at the royal settlement of Knowth, one of which produced a couple of silver coins, were the target of the plunderers in A.D. 862 and again in A.D. 934**. 2. Refuge If I have laboured this dismissal of the 'storage' explanation of Irish souterrains it is because the view is so deeply entrenched. The only serious alternative, which is completely supported by their inaccessible and complex nature, their contents (or rather lack of contents) and the contemporary Uterature, is that they were built primarily for refuge*^. It is probably due to a failure to grasp the nature of early Irish society and its dangers (and I will discuss these below) that this expla­nation is so often rejected. The constructional nature of most souterrains makes them quite ideal for internal defence by non-combatants. Creeps are invariably so narrow, or awkward, that any attempt at intrusion would be easily thwarted. The zig-zag design of many, and the building of 'drop-hole' and 'bridge' creeps has been found by investigators to completely mask the presence

Irish Souterrains: Later Iron Age Refuges 93

of occupants, even to give the impression that the souterrain is far smaller, with fewer chambers, than it actually is*'. There are also a small number** where a creep has been made so narrow that only a small child could possibly get through. The contemporary story telling of the defence of a ringfort which I have already referred to goes on to describe how the non-combatants had taken refuge in it, and were prepared to defend it. It also teUs us that the fort had a 'cave keeper'*'. Many other early stories describe the use of souterrains for refuge, but because objections have been offered to this view I feel they must be answered. The main objection is that rather than being safe refuges souterrains were traps, that the occupants could be smoked or starved out. It would in fact be impossible to cause smoke to enter a souter­rain unless all the air-vents could be discovered and blocked. Even then smoke does not go down­wards. As for starving the occupants out, this objection could be raised for any defensive structure, and requires that the attackers would be present in sufficient numbers and had sufficient time to undertake a protracted seige (several days would be necessary to starve anyone in these conditions, considerably more if they have any food and water with them). There is in any case very little evidence of seige warfare in Later Iron Age Ireland in the detailed contemporary literature. As I have indicated the average ringfort was the defended farm of a free farmer, whose wealth in portable valuables, judging by the excavation of such sites and by contemporary descriptions**, was not very great. The main items of interest to raiders would have been livestock (and perhaps occasionally humans, for sale as slaves), and there can be very little doubt that the raiders would have been from outside the tribal area in which they were raiding. For any small band of pilferers, and we can be certain that most raiding parties were small, it would have been suicidal to wait around long enough to starve the occupants of a souterrain into submission, and allow the surroun­ding farmers to rally. There are, admittedly, a number of annaUstic references to the 'smothering' (killing by asphyxia­tion) of important persons (Kings and Bishops) in souterrains by people of similar rank (A.D. 1006, A.D. 1059, A.D. 1135 for instance*''). There can be no doubt that these were semi-ritual acts, the equivalent of drowning also attested in early documents, and we may assume that the circumstances were both favourable and unusual. It is also clear from the annalistic sources that successful souterrain-plundering raids did take place, but it is equally certain that these raids involved sufficient men to overcome local defence. It is also clear from the contemporary texts that the occupants and their valuables were dug out of the souterrains**, the only safe way of forcibly entering them. The completely unpredictable zig-zag ground-plan of so many souterrains would have been an effective way of preventing ready discovery of its course. I should also stress, and it is surprising that this is overlooked by those who object to the 'refuge' explanation, that without much trouble the entrance, and even existence of a souterrain could have been concealed beneath and within the ramifications and muddle of the buildings of a working farm. I end this section with a remarkable description of the protective value of a souterrain in a typical small settlement of eight or nine European families in the American mid-west of the late 19th century*'. On receipt of the first warning of a rising by Ute Indians, the men of the settlement being away on a distant cattle-drive, the women retired every evening to a hastily provisioned

94 Richard Warner

small underground chamber. This was some 3.5 metres square, 2 metres high, with a thick earth-covered wooden roof slightly raised above the surface but impossible to fire. The chamber, large enough for twelve or fifteen people, was connected to the largest log cabin in the settlement by a narrow passage, some 4 metres long and a httle more than 1 metre high, sloping slightly up towards the cabin.

RITUAL ASPECTS I have akeady referred to the uneven nature of the distribution of souterrains, which is not re­flected by other contemporary settlements, and have suggested that a cultural or religious anti­pathy towards them may have prevented their use in some areas. I am, of course, aware of the danger of using a 'ritual' or 'religious' explanation for phenomena that seem to be otherwise illogical or inexpUcable, but in this case it might be appropriate. One of the 'three dark places of Ireland', an entrance to the underworld, was Uam Cruachain ('the cave of Cruachu') at Rath-croghan, Co. Roscommon. It is, if identified correctly, a long natural cave, or cleft, in a complex of ritual and secular earthworks, some belonging to the Later Iron Age^" (the other two 'dark places' were also natural caves). This particular cave has had its entrance made into a souterrain, though whether to enhance its ritual status, or convert it to secular use is not clear. On an island in Lough Derg, Co. Donegal, stood until Late Medieval times a small stone structure, partly under­ground and possibly a souterrain, which was known as 'St. Patrick's Purgatory' and was beUeved to be a gateway to Hell, in which visions could be had'*. A number of quite different 'ethnic' groups can be isolated within Later Iron Age Irish society, on the basis of contemporary writings'^. It might well be that some of these peoples had such a profound fear of being underground that not even the security offered by the souterrain could overcome it, and perhaps the use of Christian cross-slabs and inscribed memorial stones in the building of a number of souterrains, and even the placing of the skull at Cahercommaun, was an attempt to remove any evil otherwise associated with them. It is interesting that one of the slabs with which the souterrain section of Uam Cru­achain was built bears, in Ogham, the names of two of the 'semi-gods' who figure prominenfly in the early stratum of mythological literature.

ORIGIN AND CONNECTIONS If we regard the souterrain simply as an underground structure, accessible and non-funerary, we find that it is widespread in place and time. Restricting our enquiry only to those areas and cul­tures plausibly, or claimed to be, in contact with Ireland at a time immediately anterior to, or contemporary with, Irish soutenains we find such structures in Brittany, west Cornwall, Scotland and France'' ' . The usually tunnelled Armorican souterrains have little in common with those in Ireland, in terms of significant detail. They date, accordmg to the large quantities of pottery often found in their carefully deliberate infill and a number of radiocarbon determinations, to the middle part of the La Tene Iron Age, with no evidence of their surviving in use later than 100 B.C. They show slight regional variations, but no obvious clustering, the distribution being fairly even.

Irish Soutenains: Later Iron Age Refuges 95

The dry-stone, trench-built Cornish souterrains'' belong, as far as they can be adequately dated, to the Pre-Roman Iron Age, and into the Roman period as late as the 2nd century A.D. Quite close similarities can be found with the Irish souterrains, for instance the use of beehive chambers and their occasional siting in forts. The dry-stone, trench-built Scottish souterrains divide into a number of regional clusters, each morphologically distinct'*. They range from narrow restricted forms in the western Islands to the roomy paved, wooden roofed structures in the east. They range in date from the very Late Bronze or Early Iron Age (depending on the date of the Jarlshof roundhouses") to, perhaps, the second century A.D., although only a small number have been dated. Childe saw them all as interrelated and belived their primary function to be refuge'*. Wainwright" argued that no single explanation should be sought for all these types, and suggested that the roomy eastern ones at least were primarily for storage. Watkins*" noted however that the absence of evidence for any traffic in, or real use of, the Newmill 'roomy' souterrain seemed to go against this explanation, as did its restricted entrance. Other writers have similarly objected to the storage explanation on the general grounds of inaccessibility*' but have been stumped for an alternative, rejecting the 'refuge' explanation for no very good reason. In fact there seems to me to be no reason why all the Scottish souterrains should not have been refuges. Only the western souterrains show features comparable to those in Ireland, but lack the complexity (as do all those we have discussed so far). The French souterrains*^ seem, on the whole, to belong to the I l t h to 17th centuries, and struc­turally bear little resemblance to those in Ireland. Their real interest hes in their density in certain regions (such as on the Vienne and Charente) and their accepted function as 'passive' or 'strategic' refuges, as well as their use as hideaways for heretical sects. They also show, as we will come to again, how readily the idea of a subterranean structure was invented, and turned into reality, when required. Mahr sought *' to connect all these souterrains on the very questionable grounds that the regions in which they appeared were also noted for their megaliths, thus his 'sub-megaUthic' tradition. More recently the imprecise and misleading claim of a 'Celtic' cultural status for these same areas has led a number of writers to accept the assumption of a cultural connection between the souterrain users and thus of the diffusion of the souterrain, presumably northwards**. It is certainly true that hnks can be shown between Armorica and Cornwall at the time of the souterrains, and diffusion may well be the explanation in this case*'. But except for the most insubstantial and insignificant links** no serious claim can be upheld for any major contact between Scotland north of the rift valley, west Cornwall or Brittany on the one hand and Ireland on the other at the time of the popularity of souterrains outside Ireland. None of the well-known material and structural assemb­lages which accompany the souterrains of Brittany, Cornwall and Scotland have ever been found in Ireland. In addition the period within which souterrains can be shown to have been constructed in Ireland begins substantially later than the termination of provable souterrain use in those three other areas. As I shall argue elsewhere the Irish Later Iron Age is almost totally innovative, owing very Utile to the preceding Irish Early Iron Age*'. That is not to say, however, that aU its features are intrusive.

96 Richard Warner

though a number are (mostly from Late Roman and Early Saxon Britain). For much of the early part of the first millennium A.D. Judging by tribal origin-legends and the geographical distribution of tribal names**, there was a great deal of population movement in Ireland, with large-scale migra­tions and a high level of instability. To a smaller degree this continued for most of the millennium, and beyond, but it is generally accepted that the main tribes had attained reasonable locational stability by the 8th century A.D. By this time, and increasingly therefrom, we find the estabhsh-ment of permanent habitations. The basic habitation was, as we have seen, the single cattle-farm, often defensively enclosed, with a support area of some 20 hectares (mostly pasture)*'. It is also clear that the tribal areas were relatively small and the individual tribes almost permanently in a state of enmity with at least one neighbour. Raiding and pillaging by small bands of neighbouring tribesmen, or marauding Gall-Gaedhil, intent on cattle on the hoof, slaves, or heads seems to have been endemic'". No central provision for tribal defence has yet been demonstrated, indeed the tribal kings and aristocracy lived in precisely the same sort of single-stead habitations (ringforts, crannogs) as their people" . Protection would have been individual, ultimate security lying in the closeness of settlements (seldom more than a kilometre between neighbouring farms, judging by the present distribution) and the blood ties and ethnic and political unity of those neighbours. Clearly, in such circumstances, the ringfort, as a primary defence around the farm buildings and stockyard, was a simple and eminently practical solution, as was the souterrain as a secondary level of defence for non-combatants. And I see both as an Irish invention, a response to a parti­cular problem within a particular socio-economic structure. Thomas has shown'^ that a very simi­lar economy based on cattle and holdings of about 20 hectares around individual farms existed in the Early Iron Age in Cornwall, and it is no surprise that ringfort-like enclosures are a common feature of that culture, as are (though limited in distribution) souterrains. It would therefore seem to me that the key to the distribution of souterrains from Brittany to Scotland lies not in any notional cultural link, but in similar socio-economic enviroimients inviting similar inventive responses.

Irish Souterrains: Later Iron Age Refuges 97

NOTES 1) S.P. 6 Riordain, Antiquities of the Irish Coun­

tryside (1964), esp. 27-34 (and revision by R. de Valera, 1979, esp. 65-73).

2) R.B. Warner, The Irish Souterrains and their Background. In: Subterranean Britain, Ed. H. Crawford (1977) 100-144.

3) Warner (cf.n. 2) fig. 4.1. 4) Warner (cf. n. 2) 101-3, figs 4.2,4.3. 5) For instance, D.C. Twohig, Souterrains at

Sheepwalk and Ballirisode, Co. Cork. J. Cork Hist, and Archaeol. Soc. 78, 1973, 35-39, fig. p. 37.

6) B.T. SomervUle, Descriptive Account of the Curraghcrowly Souterrain, Ballineen. J. Cork Hist, and Archaeol. Soc. 35, 1930, 1-16; May and Cooper (cf. n. 65).

7) Warner (cf. n. 2) 104-11, figs 4.4 to 4.7. 8) For instance, R. O Floinn, A Souterrain at

Termonfeckin, Co. Louth. Co. Louth Archaeol. J. 19,1978,128-30.

9) E. Ryime, Souterrain at Donaghmore, Co. Louth. Co. Louth Archaeol. J. 14, 1959, 148-53.

10) For instance Knockdhu, Co. Antrim, in H.C. Lawlor, Some Notes on the Investigation of DweUing Places of Prehistoric Man in N.E. Ire­land. Proc. Belfast Nat. Hist, and Phil. Soc. 1917,31-61.

11) For instance, R.A.S. MacaUster and R.Ll. Praeger, Report on the Excavation of Uisneach. Proc. Royal Irish Acad. 38c, 1929, 69-127, pi. 8, redrawn Warner (cf. n. 2) fig. 4.15.

12) For instance, M. Stenberger, A Ring-Fort at Raheennamadra, Knocklong, Co. Limerick. Proc. Royal Irish Acad. 65c, 1966, 37-54.

13) S.P. 6 Riordain and P.J. Hartnett, The Exca­vation of Ballycatteen Fort, Co. Cork. Proc. Royal Irish Acad. 49c, 1943,1-43.

14) Warner (cf. n. 2) 121-22. 15) P. Power, Find of Bronze Celts at Aghadown,

Co. Cork. J. Royal Soc. Antiq. Ireland 56, 1926, 57-58.

16) M.J. Rowlands, The Production and Distribu­tion of Metalwork in the Middle Bronze Age. British Archaeol. Reports 31,1976, appendix J.

17) J.S. Jackson, Metallic Ores in Irish Prehistory: Copper and Tin. In: The Origins of Metallurgy in Atlantic Europe (Proc. 5th Atlantic Colloq., 1980) 107-25. Map 1, area of M4-7.

18) 6 Riordain, Excavations at Cush, Co. Limerick.

Proc. Royal Irish Acad. 45c, 1940, 83-181. 19) 6 Riordain (cf. n. 18) 110-12,116. 20) R.M. Kavanagh, Collared and Cordoned Ciner­

ary Urns in Ireland. Proc. Royal Irish Acad. 76c, 1976, 293-403, 366 no. 28.

21) For instance M. Herity and G. Eogan, Ireland in Prehistory (1977) 223, 230-32, esp. 231.

22) CJ. Lynn, The Datmg of Raths: An Orthodox View. Ulster J. Arch. 38, 1975, 45-47. R.B. Warner, The Beginnings of Fortification in Later Iron Age Ireland. BuU. Ulster Place Name Soc. 3 ,1981,45-52.

23) D.B. Harden, Glass Vessels in Britain and Ire­land, A.D. 400-1000. In: Dark Age Britain, Ed. D.B. Harden (1956) 132-67, 154; photo in Warner (cf. n. 2) fig. 4.16.

24) M. Dolley, The Anglo-Saxon Pennies from the Upper Souterrain at Knowth. British Numis. J. 38, 1969, 16-21. For the site and date: Eogan, Report on the Excavations of some Passage Graves, Unprotected Inhumation Burials and a Settlement Site at Knowth, Co. Meath. Proc. Royal Irish Acad. 74c, 1974, 11-112, esp. 87-111.

25) Information from the National Museum, Dublin.

26) Stenberger (cf. n. 12). 27) S.P. 6 Riordain and M. MacDermott, Excava­

tion of a Ring-Fort at Letterkeen, Co. Mayo. Proc. Royal Irish Acad. 54c, 1952, 89-119.

28) M. Ryan, Native Pottery in Early Historic Ire­land. Proc. Royal Irish Acad. 73,1973, 619-45.

29) For instance, Macalister, Corpus Inscriptionum Insularum Celticarum 1, 1945, 267, 297; 2, 1949,121.

30) K. Hughes, Early Christian Ireland: Introduc­tion to the Sources (1972) esp. chapters 4, 5.

31) A.T. Lucas, Souterrains: the Literary Evidence. Bealoideas41, 1973,165-91.

32) C. Thomas, J. Royal Inst. Cornwall 6, 1972, 251-74.

33) A crude map will be found in Lucas (cf. n. 31). 34) See for instance the map in E.M. Jope et al., An

Archaeological Survey of County Down (1966) fig. 72.1; souterrains are confined to the south of the county.

35) For instance E. Watson, Prehistoric Sites in South Antrim. Ulster J. Arch. 3,1940,142-51, fig. 2.

36) Ryan (cf. n. 28) fig. 4.

98 Richard Warner 37) Lucas (cf. n. 31) 166. 38) G. Coffey, New Grange and other Incised Tu­

muli in Ireland (1912, repr. 1977), compare fig. 2 with fig. on 15.

39) A. Mahr, New Aspects and Problems in Irish Prehistory. Proc. Prehist. Soc. 3, 1937, 262-436,386.

40) V.B. Proudfoot, The Economy of the Irish Rath. Medieval Arch. 5, 1961, 94-122; 6 Rior­dain (cf. n. 1) 1-12.

41) Cf.n. 11. 42) 6 Riordain (cf. n. 18). 43) E.R. Norman and J.K. St. Joseph, The Early

Development of Irish Society (1969) 90-121. 44) L. de Paor, The Viking Towns of Ireland. In:

Proc. 7th Viking Congr., Eds B. Ahnqvist and D. Greene (1976) 29-37, 29.

45) Lucas (cf. n. 31) 172-73 gives a short list. 46) For instance D.M. Waterman, A Marshland

Habitation Site near Lame, Co. Antrim. Ulster J. Arch. 34, 1971, 65-76. Waterman, The Ex­cavation of a House and Souterrain at Craig HiU, Co. Antrim. Ulster J. Arch. 19, 1956, 87-91.

47) For instance Macalister (cf. n. 11), Waterman (cf.n. 46, Craig Hill).

48) S.P. 6 Riordain and S. Foy, The Excavation of Leacanabuaile Stone Fort, near Caherciveen, Co. Kerry. J. Cork Hist, and Archaeol. Soc. 46, 1941, 85-99, redrawn Warner (cf. n. 2) fig. 4.11.

49) G. Du Noyer, On the Remains of Ancient Stone-built Fortresses and Habitations Occur­ring to the West of Dingle, Co. Kerry. Archaeol. J. 15,1858, 1-24.

50) Lyrm, BaOywee. In: Excavations 1974, Ed. T. Delaney (1975) 4-6.

51) H. O'N. Hencken, Cahercommaun: A Stone Fort in Co. Clare. Extra vol. of Roy. Soc. Antiqs Ireland (1938).

52) Jopeetal. (cf.n. 34) 185-96. 53) Eogan (cf. n. 24). 54) E. Evans, Rath and Souterrain at Shaneen Park,

Belfast. Ulster J. Arch. 13, 1950,6-27. 55) T.F. Murphy and C. 6 Cuileanain, A Ring-Fort

at Oldcourt, Co. Cork. J. Cork Hist, and Arch-aeoL Soc. 66,1961,79-92.

56) Hencken(cf. n. 51). 57) O Riordain, Lough Gur Excavations: Carraig

Aille and the Spectacles. Proc. Royal Irish Acad. 52c, 1949, 39-111,62-64.

58) Lucas (cf.n. 31) 175. 59) Lucas (cf. n. 31) 179.

60) Annals of Ulster. Ed. W. Hennessy and B. Mac-Carthy (1887-1901) 1, 373; Annals of Inis-faUen,Ed. S. Mac Airt (1951) 134.

61) Annals of Ulster 1, 45; Annals of the Four Masters, Ed. J. O'Donovan (1856) 2,631.

62) Rynne (cf. n. 9) is one of the few Irish archaeo­logists to have subscribed to this view in print.

63) Rynne (cf. n. 9) 152. 64) For instance J. Waddell, Notes on Some Kerry

Souterrains. J. Kerry Archaeol. and Hist. Soc. 3, 1970,15-18,16. Ryan, A Souterrain in Keal-duff Upper Townland, Glenbeigh. J. Kerry Archaeol. and Hist. Soc. 9, 1976, 5-10, 10. A.McL. May and D.C. Cooper, Rock-Hewn Sou-terrain, Ballintemple, Garvagh, County London­derry. Ulster J. Arch. 2, 1939, 82-97, this is the only possible explanation of opening 'H' on their fig. 1 as 'D' was blocked off.

65) Lucas(cf.n. 31) 185,187. 66) Proudfoot (cf. n. 40); L. and M. de Paor, Early

Christian Ireland (1958) chapter 3. 67) Annals of the Four Masters 2, 759; Annals of

Ulster 2, 5; Annals of Loch Ce, Ed. W. Hennes­sy 1 (1871) 135.

68) Lucas (cf. n. 31) 171,184. 69) W. Thayer, Marvels of the New West (1888) 259. 70) S. Ferguson, Account of Ogham Inscriptions in

the Cave at Rathcroghan, County of Roscom­mon. Proc. Royal Irish Acad. 9, 1867, 160-70; Lucas(cf. n. 31) 188-89.

71) Lucas(cf.n. 31)189-91. 72) T.F. O'Rahilly, Early Irish History and Mytho­

logy (1946) passim. 73) A general survey of souterrains will be found in

no. 2 of the Document series of Archeologia (1973) where all the regions with which I deal here are discussed.

74) P.-R. Giot, Les Souterrains Armoricains de I'age du Fer. Ann. de Bretagne 67,1960,45-65; Giot et al., Protohistoire de la Bretagne (1979) 292-300.

75) Hencken, Archaeology of Cornwall and ScUly (1932); E. Clark, Cornish Fogous (1961).

76) F.T. Wainwright, Souterrains in Scotland. Antiquity 27, 1953, 219-32; Wainwright, The Souterrains of Southern Pictland (1963); D. Brothwell, On a Mycoform Stone Structure in Orkney, and its Relevance to Possible Fur­ther Interpretations of So-called Souterrains. Bull. London Inst. Arch. 14,1977, 179-89.

77) J.R.C. Hamilton, Excavations at Jarlshof, Shet­land (1956) 32-39.

Irish Souteirains: Later Iron Age Refuges 99

78) V.G. Chflde, Prehistoric Communities of the British Isles (1940) 160.

79) Wainwright 1963 (cf.n. 76 )13 . 80) T. Watkins, The NewmiU Souterrain. Current

Arch. 6 ,1979, 205-9, esp. 208,9. 81) For instance, Brothwell (cf. n. 76) 187,8.

J.X.W.P. Corcoran, The Souterrain at Rosal, Strath Naver, Sutherland. Proc. Soc. Antiqs Scotland 100, 1968,114-18,117.

82) A. Blanchet, Les Souterrains-Refuges de la France (1923); P. Piboule, Les Souterrains Amenages du Chatelleraudais. Arch. Medievale 1,1971,241-60.

83) Mahr (cf. n. 39). 84) C. Thomas, Souterrains in the Irish Sea Pro­

vince: A Note. In: The Iron Age in the Irish Sea Province, Ed. C. Thomas (1972) 75-78; Childe

(cf.n. 78) 261,263. 85) Thomas, The Character and Origins of Roman

Dumnonia. In: Rural Settlement in Roman Britain, Ed. C. Thomas (1966) 74-98.

86) Waddell, From Kermaria to Turoe. In: Ireland 800 B.C. to A.D. 800, Ed. B. Scott (1982).

87) For instance, Warner (cf. n. 22). 88) O'Rahilly (cf. n. 72). 89) Proudfoot (cf. n. 40). 90) dePaor(cf.n. 66) 77,158. 91) Warner, The Excavations at Clogher and their

Context. Clogher Record 8, 1973, 5-12; Hen­cken, Lagore Crannog: An Irish Royal Resi­dence of the Seventh to Tenth Centuries. Proc. Royal Irish Acad. 53c, 1950, 1-247; Eogan (cf. n. 24).

92) Thomas (cf. n. 85) 96.