More information : (SO 761374) Hillfort (NR) (1)
Limited excavations in 1924 by I.T Hughes. Hut hollows and a mound
yielded pottery, charcoal, flints and iron, the finds being on the
surface below the mound. A hut was stone floored and a possible
guard-hut had a stone platform. Sections of the ramparts showed
that there were originally two banks (the inner one revetted with
stone and partly built with material from inside the camp) with a
ditch between. When the ditch filled up, a paved way was laid along
it. Finds here included pottery, charcoal, burnt clay and quartz
and clay sling-stones.
All the pottery was of La Tene type, soapy, some with fingerprint
and incised-line ornament. No Roman finds. (2)
The excavations "established here an Iron Age B date for the rubble-built main rampart". (3)
Two original entrances: the northern entrance is approached by a
curved trackway, while the inner rampart on the east is turned
slightly inward: the western entrance has both the inner ramparts
inturned. A slight rampart runs from X-Y, with traces of an
internal ditch, and curves round towards the Red Earl's Dyke (q.v.
SO 73 NE 9). "It is uncertain if this or both banks formed an
outer enclosure, but in 1875, 40 hut-sites could be traced within
the area. The small stream running down the valley has been dammed
at 3 points to form reservoirs; once at the spring just within the
inner enclosure to the west of the south entrance, and twice
within the outer area. "(See Plan). (4)
A major hill-fort. Published survey (25") revised. A summary of the
1967 excavations will be forwarded to the Archaeology Division by
the Malvern Hills Archaeological Committee. (5)
Excavations 1965-1969 have shown that the Camp was persistently
fortified and permanently occupied over a long period. The earliest
phase appears to have begun c 300 BC and the occupation is of IA `B'.
(6)
Midsummer Hill Camp is a mainly univallate hillfort enclosing 22 acres (8.9h). Excavations from 1965 till 1970 were aimed at elucidating the history of the South Gate and the nature of any internal occupation. At the gateway seventeen phases were
recognised involving guard-rooms in the early stages and bridges in the latest ones. The camp was established in the early 5th century BC and occupied until the Roman Conquest by Ostorius Scapula in AD 48. The site was occupied as a village of 1500-2000 people from the beginning and where the terrain permitted, oblong four-post buildings were set in lines; but most of the interior has gradients of 1 in 3 and 1 in 4, and even these slopes were utilized. Small terraces cut into the hillside are thought to be emplacements for huts and more than 200 have been plotted. Excavated huts show repeated post replacement, involving between four and six phases in the course of the 500 years of occupation. Some iron and bronze was worked on the site although the furnaces have not been located. Finds of bronze are few. There was earlier, Beaker activity on the hill; and Medieval or Post-Medieval ploughing. (7,8)
SO 760 375. Midsummer Hill. Listed in gazetteer as a multivallate hillfort covering 7.0ha. (9)
Midsummer and Hollybush Hills. Description with plan. (10)
The hillfort has been mapped as part of the Malvern Hills AONB NMP. (11-12)
The earliest find from the site is a reused Neolithic stone axe (8) but excavations by Hughes (2) as well as Stanford produced evidence of other prehistoric flint work (8). Although much of this was made on flint pebbles it?s presence is nevertheless quite significant in view of the distance from flint gravel sources. Stanford's excavations on the Hollybush col revealed some of the rare pre-hillfort deposits to have escaped later prehistoric disturbance. These contained twelve Beaker sherds all possibly associated with small hollows and therefore felt to represent domestic rather than burial activity (8). No evidence of the presence of prehistoric barrows was encountered during the RCHME survey, and suggestions that mounds to the north and south of the pillow mound are barrows could not be confirmed. That to the south has the best surface evidence. It may be that it was the spring that provided the focus of this early attention, particularly positioned as it is amongst dramatic hills.
Given this activity, the likelihood of pre-hillfort earthworks and indeed earlier enclosures existing might be considered high. Unfortunately aside from the damage to the natural contours done by the extensive terracing, the steep slopes ensure that erosion and hillwash here must have been considerable. Nevertheless, many of the isolated depressions could be of an earlier period, and any one of the terraces, particularly the higher ones, may represent the line of a former enclosure around Midsummer Hill.
Certainly a former line of the hillfort enclosure itself can be traced on the surface. Not only does this underlie the main hillfort defences in a number of places, but chronology is assisted by the presence of the Shire Ditch, one element of which appears to underlie and the other overlie the hillfort defences. Assuming the element that underlies the hillfort to be Late Bronze Age, analogous to linear ditches in Wessex, the earliest phase of hillfort enclosure might have been constructed as early as the Late Bronze Age. However, while Stanford's excavations at the south entrance revealed seventeen distinct phases (8), none produced pottery or other artefact evidence of such an early build.
Neither ditch nor bank is of great proportions, although the manner in which the slope has been utilised gives the impression of a construction of more massive bulk. The excavations by Hughes (2) indicates that construction of the Midsummer Hill counterscarp at least was of two phases or more. The paving revealed within the ditch by Hughes is intriguing. If not the result of tumbled bank revetment (and the order revealed in the published plans would suggest that is not the case), there would appear to be some unexplained use of the ditch. Hughes suggested that the ditch was periodically used as a trackway. This might be thought to negate use of the ditch for defence.
With one rampart terminal inturned, the northernmost entrance has been made difficult to access, while the southern has been provided with an additional breastwork (now partly obscured by ponds). While this might be seen as providing added protection for the more vulnerable gate, it is nevertheless overlooked from the valley sides and its elaboration might be considered to be for display as much as for defence. The refined approach channels activity and enables a desired processional route to be observed, in a similar manner to the entrances to banjo enclosures. As noted above Stanford claimed some seventeen phases of activity at the south entrance, although how much these phases reflect general reconstruction or overhaul of the defences as a whole and how much they reflect changes at the entrance is not clear. Early plans of the hillfort (in authy 2 and 8) indicate that there may have been a third entrance at the southernmost tip of the Hollybush Hill ramparts. Hughes depicts the rampart terminals here as slightly offset and with a simple gap in the counterscarp, although Lines illustrates the entrance apparently cutting through the ramparts suggesting that it might be a later addition. Unfortunately, as this area has now been completely quarried away, these early plans are all that we have to go on.
Why the hillfort should be placed in such close proximity to that on Herefordshire Beacon (there is less than 2km between them) is equally unclear. Location on the summit of the Malvern chain might be considered of strategic advantage, particularly where it is possible to control routes through the hills. Herefordshire Beacon controls the Wynds Point pass, and Midsummer Hill the Gullet and Hollybush passes. However, such close proximity is thought unlikely to be purely for reasons of regional defence and there are other hillforts nearby, Wall Hills Camp, Haffield Camp and possibly Eastnor, that will have influenced the situation. Like Midsummer Hill Camp, the British Camp on Herefordshire Beacon developed in a series of phases (SO 74 SE 3), each of which may have overlapped with contemporary activity at Midsummer Hill Camp, and as likely to be complementary of it as antagonistic. Midsummer Hill Camp itself is quite unusual in not only encompassing two prominent hilltops but also the dramatic ravine between them. Whereas the British Camp sits proud on the skyline and boasts status, Midsummer Hill Camp is more ephemeral, and despite comprising highly visible ramparts, is nevertheless inward looking, its focal attraction being the internal sheltered valley and the spring that it harbours.
The extent to which settlement survived independently of the hillfort is unclear, as is the thorny question of which came first. Terraces and huts cannot be traced elsewhere on the hilltops, and while many of the unexcavated stances, particularly those unaffected by terracing on the summit of Midsummer Hill, could be of earlier date, the earthworks suggest that for the most part settlement developed within the confines of the hillfort. In terms of the `British Town' little confirmation could be added to indications that the lower part of the ravine held settlement. This is not to say that it didn't exist, for situated at the foot of steep slopes the area is likely to have received much colluviation and it corresponds with the area demarcated as `garden' on the Eastnor Tithe map. In addition, most of the area has been affected by quarrying activities during the last century. A few ephemeral platforms were traced outside the hillfort defences during the present survey and as noted above, these appear to be later than the hillfort ramparts, but they are seen as extra-mural from the hillfort rather than as part of a major external settlement. The earthwork evidence indicates that, in general, settlement was indeed constrained by the hillfort defences, and at no point can settlement, terraces, or lines of platforms, be seen to underlie the defences.
The interior of the hillfort is densely packed with hut stances, most of which are placed in a regular manner along the contours. Relatively few are situated within the internal quarry ditch, which may indicate that it was used for rampart repair, or was kept clear for other purposes. Alternatively, silting from the interior may have obscured features here, and it is worth noting, for example, the presence of four-post structures set just inside the rampart at Grimethorpe (Stead 1968). Many huts at Midsummer Hill Camp are set along the lip of the internal quarry ditch and from here extend across the interior. Rarely is there surface evidence to suggest that they overlap or were chronologically distinct although a few at the south end of Midsummer Hill may come into this category. Equally the terraces upon which many structures were built are aligned along contours. On the north-east slopes of Midsummer Hill the terraces observe a slightly different course, the two alignments abutting at the point where the modern footpath between the entrances traverses the site. The footpath is partly terraced and may reflect the 'ghost? of a convenient route across property units, which were then reorganised in response.
Why the settlement is positioned in such a marginal location is unclear. The site is extremely exposed and at certain times of year it is difficult to stand against the prevailing south-westerlies, and yet the location of many hut stances means that they receive this head on. If we take the surface evidence at face value as others have done and assume that the gridiron plan revealed by Stanford, and the construction of terraces for hut stances, implies an intensively utilised environment, it might be worth enquiring about the lack of internal route ways. This lack of access ways was observed by H H Lines in 1870. Despite preservation of the slightest hut depression, not a single original hollowed trackway remains . This is true for the external slopes too. Assuming constant movement to and from the valleys below the hills, whether to tend fields on the lower ground around the base of the Malvern Hills and along the Avon terraces, or to graze stock, or quarry stone, the surrounding slopes might be expected to be heavily engraved with trackways. Equally there is no indication internally of stockyards or areas where stock can be kept, or milked. Within the hillfort, paraphernalia usually associated with the farmyard is missing and for an intensively settled interior, there is little indication of the wear and tear of everyday activity.
Although cautious, to Stanford most building stances represented domestic huts, and from the total of 244 huts, half of which were considered dwellings, he was able to suggest that between 1500 and 2200 people lived within the fort (8). Almost 480 stances were recorded during the present survey, and if we allow for structures not revealed by surface traces, such as those excavated on the Hollybush col, the kind of population envisaged by Stanford can only be confirmed rather than reduced. Certainly there is a good case for some of the stances being domestic, particularly those on the terraces where it is possible to envisage domestic units. However, many are much too small for domestic occupation, and those cut into hard, impervious, natural rock can hardly have been for subterranean storage. Similarly, Stanford interpreted the four-post structures arranged in grid-iron fashion on the Hollybush col as a mixture of domestic buildings and storehouses (8). Thirty years on, there might be a tendency to assume that all were granaries. However, nine contained potsherds and three hearths, and of thirty-one structures there considered to represent four phases of activity, only seven produced any evidence of grain. Within the sixteen structures burnt down, where optimum survival conditions might apply, grain was only found in five. Divorced spatially from circular hut stances, in a similar manner to some other hillforts, there is some implication that if used for storage these four-post structures are likely to be communal in nature, and if so, larger buildings could easily have been used to save overall space. Although some were quite large (up to 4.6 by 3.4m) none remotely resemble even small tithe barns of the medieval period for example. While there may have been some symbolic or social importance attached to the storage of grain above tools or other commodities, there remains the probability that soils and climate around the Malverns was more receptive to root crops and to pastoral farming, and if indeed storehouses, the four-post structures may well have held chickens, cheese or other produce.
Other interpretations can be placed on both stances and the excavated four-post structures and Stanford himself recognised that some depressions may even be tree throw holes. Found singly elsewhere in the Malverns, small depressions or scoops are often considered to be early quarries, and scoops or levelled areas for modern seats, benches and similar features are often indistinguishable from individual hut stances. Here, however, the sheer number, the organised arrangements, and associations indicate that almost all are archaeologically significant. Where such stances have been excavated elsewhere little structural evidence appears to be present to supplement the investment of cutting the platform in solid rock. The lack of drainage channels, in particular, might be of concern. Equally, while associated, domestic detritus is not present in great quantities and the evidence for hearths, though again present, is particularly slight given such an inhospitable environment. The evidence encourages an impression of single events rather than permanent occupation.
How such stances are interpreted, however, remains of importance. At Midsummer Hill Camp their closely packed nature does suggest intense activity, or at least competition for space, and many must have been contemporary or near contemporary. Stanford?s excavations on the Hollybush col indicate that most of the four-post structures were continually replaced in the same position rather than reconstructed on fresh sites elsewhere within the hillfort and the possibility that some of the examples here may have been treated in similar fashion cannot be ignored. The plan so engraved is of a well-established blueprint; the construction of the terraces has determined the form and nature of settlement within the fort. No mean effort has been invested in providing areas to construct buildings upon, although why this should be and whether this concentration within the defences was for protection is less than clear. The slopes around many structures, particularly on Midsummer Hill are so steep that they would be almost unusable in an everyday domestic or agricultural context. Even so domestic functions are generally envisaged for such stances and the explanation, at least in part, put down to transhumance. At Midsummer Hill Camp, this would probably rule the site out as an agricultural base, as investment in crops is likely to discourage absence. Equally, if we were to accept that the large number of buildings reflects the size of the population, then enormous numbers must have been moving around the countryside.
Recent interpretations of prehistoric landscape lead us to question whether such remains indeed represent settlement. Where hillforts developed in response to social and ritual requirements of the Late Bronze Age, use might be more akin to the various functions formerly fulfilled by Neolithic causewayed enclosures: meeting places, fairs, markets, etc, all carried out with ceremonial rather than defensive overtones. Given such a role, the building stances might be considered in a different light, performing a variety of transient, but traditional, social, agricultural exchange, ceremonial, and ritual functions. Given problems of identification of Iron Age burial practices, some of the stances and even some of the four-post structures could be excarnation platforms. Similarly, the ludicrously small 'guard chamber?, in the north positioned behind the rampart, is badly sited as a lookout post or gatekeeper lodge, but would be fine for the positioning of an icon, idol or statue. This must be close to Hughes site no 3 (2) which is depicted on his plan as being some feet west of the north entrance, but which he felt represented a guard chamber. A depression in the east rampart terminal might be a similar feature. Those at the southern entrance, revealed by excavations, were situated in the rampart terminals either side of the entrance corridor, and while well placed for a gatekeeper are less so for a guard or lookout post.
Although the hillfort defences encompass the summits, the focus of activity appears to have been in the valley between the two hills. This is not surprising, for not only is it sheltered from the biting winds, but there is a superb source of water. The water in the Malverns has certainly been respected as particularly pure during historical and recent times and some springs particularly revered and given names, such as St Annes Well. The position of the Hollybush spring, hidden in the declivity of two striking hills, might only add to its importance and enhance its position as a site of some symbolic value. The Malvern summits themselves can be seen from 20km or more, and provide a dramatic contrast to the surrounding low ground, particularly of the Severn valley. The symbolic importance of such natural places, particularly of dramatic hills on the one hand and springs on the other, as providing interfaces with the spirit worlds, is widely known from ethnographical writing and their role within social and ritual landscapes now discussed quite extensively within archaeology.
The ponds in the valley, variously referred to as cisterns or tanks of British or Romano-British date that serviced the `Town', utilise and adapt the hillfort defences in a manner that would render the latter unusable). This and their sharp profile encourage the view that they are much later in date than the settlement. They may have been for the use of stock, but the flight of at least six examples (in addition to the four bays recorded here, two others are depicted downstream on the 1st edition 25" map of 1886) indicates that they may have served an industrial purpose, or perhaps even have been part of the `garden' mentioned on the Eastnor Tithe Apportionment.
The nature of medieval and later activity is elusive. Finds of medieval iron knives and potsherds by Stanford (8) from around the southern entrance indicate that the springs were still an attraction. Some of the terraces, particularly the lower examples in the north of the Midsummer Hill slopes which are almost devoid of building traces appear to have been cultivated and excavations on the Hollybush col revealed striations in the soil interpreted by Stanford as ploughmarks (8). Although slender there is perhaps enough evidence to suggest that a smallholding occupied the area behind the south entrance or the nearby slopes, which may have been responsible for the medieval finds as well as the soil disturbance that Stanford encountered on the Hollybush terraces. It may also have been responsible for some of the cross divisions placed within the Hollybush quarry ditch, though these are equally difficult to explain. They appear to provide stall-like units though perhaps there are too many to suggest that they form the cattle or sheep pens of a small farm. The use of hills (e.g. Tan Hill, Wilts: Weyhill, Hants) and of hillforts (e.g. Yarnbury, Wilts) for medieval fairs readily springs to mind and the place-names here might support suggestion of such an event on certain holy days.
There is little doubt that the long mound (SO 73 NE 10) placed centrally on Hollybush Hill was for the propagation of rabbits, as both surface and excavated evidence points to that. If the hilltop were so used the ramparts would demarcate a ready-made conygar. The Hollybush ramparts are very denuded but traces of a surmounting bank are visible in places. If hedged it may have been this as much as the existence of a county boundary that was responsible for the line of trees that remains today. Only the position of the warrener's dwelling is unknown. (13)
Bronze Age material recorded separately under SO 73 NE 13. (14)
|