Issue no 36, July 1998

Tom Williamson points out some of the historic features to be found in Georgian country parks

It is sometimes imagined that the designers of 18th century parks around great country houses created entirely new landscapes, sweeping away all traces of features that had existed before.

The truth is, rather, that with a keen eye we can often still see clues to the earlier history of parkland, surviving, for example, as earthworks. Archaeologists are also now beginning to detect some common patterns in the layout of parks which can often shed valuable new light on the attitudes of the Georgian gentry.

The landscape park was not, as is sometimes suggested, an entirely new invention of the 18th century. Indeed, large-scale aesthetic landscaping was practised in the Middle Ages, when castles like Bodiam in Kent were provided with elaborately landscaped grounds, featuring areas of ornamental water and decorative tree-planting.

The origins of the landscape park lie, in part, in the deer parks of the Middle Ages. Originally valued primarily as venison farms and hunting reserves, these grew steadily in aesthetic importance and, by the later Middle Ages, had become a normal adjunct to a lordly residence. In the 15th and 16th centuries the density of trees within them was reduced, vistas were opened up and, especially from the middle of the 17th century, avenues were extensively planted.

What was new in the middle decades of the 18th century was not so much the park itself but the fact that it now became the principal setting for the house. Walled gardens, terraces, and geometric clutter were removed from the main façades, although gardens - often consisting of meandering gravel paths through grass and shrubbery - were usually maintained to one side. The orchards, fish ponds and nut trees which had formerly graced the country house, together with such functional features as barns and farm yards, were likewise generally swept away, and kitchen gardens removed to some more hidden location and screened from sight by a strip of shrubbery.

The ideal was now for the mansion to stand in the midst of open, irregularly-planted parkland, with the turf appearing to flow uninterrupted to its walls - although in reality a sunk fence or ha-ha usually separated the mown ground of the lawn from the grazed grass of the park. Only in the early 19th century did gardens begin to reappear between house and park.

Many landscape parks developed directly from earlier deer parks, although these received much new planting in the form of perimeter belts and scattered clumps: and where the terrain permitted, a lake of suitably serpentine, naturalistic form was provided, sometimes through the conversion of an earlier fish pond complex. But most parks were laid out at the expense of gardens and farmland. Lancelot `Capability' Brown - the designer we normally associate with the new style - worked on perhaps 180 landscapes during his long career but this represents a tiny minority of the total and there were many other designers. Some had national practices, but many served local markets. A number of landscape parks were simply designed by the owner, the estate gardener, or both.

So what should we look for when visiting the grounds of a country house? Anyone with an archaeological eye should look out for earthworks, which generally fall into two main categories. Firstly, there are those associated with the agricultural landscape which the park replaced. Estate accounts frequently record payments to labourers for grubbing out hedges and the like, but earthworks of field boundaries often survive, in varying states of preservation. These are important because they represent the raw material with which the park-makers worked, and can thus help us understand some of the design decisions taken by owners and landscapers.

Earthwork banks are often associated with ancient pollarded trees, which once formed parts of hedgerows which were retained when the boundary was removed and the new landscape laid out. In many parks, a high proportion of the trees pre-date the creation of the park itself. Hollow ways are also often encountered, for roads and footpaths were frequently closed when parks were created (a procedure made easier by the passing of the Turnpike Act 1773). A range of other features - mill mounds, ridge and furrow, and so on - can also be found. Such survivals are of particular importance in the arable east of England, where parks are often the only areas of permanent pasture in a parish.

It is often, however, unclear whether such remains represent features still existing at the time the park was created, or whether they had long survived only as earthworks in paddocks and pastures around the manor house. Settlement remains are quite commonly found - Mike Hughes, former county archaeologist in Hampshire, has suggested that as many as half the landscape parks in the county contain settlement evidence. Yet while many villages were indeed cleared to make way for parks, especially in the Midlands, such remains often represent places which had disappeared long before. In Norfolk, for example, between a quarter and a third of known deserted settlements lie within areas of 18th or early 19th century parks: but fewer than one in ten were abandoned as a result of the park's creation. Parks were created in parishes owned by large estates and such dominance often resulted from drastic contraction of settlement in late medieval times. Moreover, areas of desertion often provided convenient gaps in which parks could be established.

The second category of parkland earthworks relates to the `aesthetic landscape' and includes the remains of formal gardens demolished when the park was laid out (although these were often very systematically flattened), and earthworks associated with 18th - century parkmaking - cuttings made to open up vistas and the like. So careful were 18th century landscapers, so adept at their art, that such earthworks are often in effect invisible and easily taken for natural landforms. At Chatsworth in Derbyshire, for example, Brown and his associates were paid vast amounts for moving earth, yet a recent survey by the archaeologist John Barnett revealed few lacunae in the spread of ridge and furrow and other earthworks within the park which could indicate where landforms had been altered, except in the area immediately to the east of the river Derwent. Here the ground has been carefully graded, on a massive scale, in order to open up views of the river from the house, although this is not immediately apparent in the naturalistic lie of the land.

It can also be interesting to note how parks were laid out. It is noteworthy, for example, that while kitchen gardens were generally kept from view, stables were often large and proudly displayed - sometimes, as at Houghton in Norfolk, rivalling in grandeur the mansion itself. Thoroughbred horses were the élite status symbol par excellence in the 18th century, and parks provided excellent spaces for fast riding.

The location of the house within the park is also worthy of attention. Where the park is large, the mansion usually stands towards its centre. Where the park is small, it is normally placed towards the northern boundary, so that the owner could enjoy the most extensive views from the principal living rooms on the warm, south-facing side of the house.

These landscapes have much to tell us about the attitudes and priorities of their owners. The park was, above all, a landscape of seclusion - belts of trees, entrance lodges, closed roads and shifted settlements attest to that clearly enough. The park also signalled the owner's divorce from active involvement in useful production at a time when, with an expanding economy and a growing culture of consumerism, superior resources of production (fish ponds, kitchen grounds, orchards) were no longer a badge of pride.

Landscape parks were also a manifestation of changing social relations among the propertied. The expansion of the economy associated with the consumer revolution of the mid-18th century brought growing uncertainties over the very definition of gentility, as the middle classes grew in numbers and wealth and themselves acquired gardens of sophistication. In an uncertain world, Brown's landscape style, which consciously downgraded the importance of the garden, but affirmed that of the park, firmly asserted the traditional status of landowners. Only they possessed the raw materials necessary for a park's creation - land in abundance.

Dr Tom Williamson is a Lecturer in Landscape Archaeology at the University of East Anglia

Seeking the origins of bronze tools

The earliest metal goods probably came to Britain from Ireland. Paul Budd reports

Squeezing through the labyrinth of tortuous passages carved out of the solid rock at depths of 70 metres or more, it is difficult not to spare a thought for Britain's prehistoric copper miners. Some of the tunnels beneath the Great Orme's Head, near Llandudno in North Wales, are so small that they could only have accommodated children.

The experience is fascinating and the conditions brought vividly to light. Crawling on hands and knees by the light of a dimly burning taper clenched between the teeth; huddled in the gloom and pounding at the fire-softened rock with a cobblestone hammer; the work must have been excruciating. And yet, 20 years ago this and other astonishing evidence for the earliest copper mining in Britain did not exist and the passages lay undiscovered.

Today, we are entering a new phase of research on Britain's earliest copper mines. Much of the excavation and recording has taken place, telling us when and how the mining and ore processing was carried out and, to some extent, about the people who did it; but where did all the copper go? A long-standing objective in archaeometallurgy has been to try to link Bronze Age metal tools and weapons to their sources. Now, perhaps for the first time, scientific techniques are beginning to tell us something about this vital key to understanding the organisation of prehistoric metal production and exchange.

Until the early 1980s only one prehistoric copper mining site was known in the British Isles. The site, Mount Gabriel on the Mizen peninsular in the far south-west of Ireland, was simply a cluster of primitive opencast workings and shallow galleries dug into the hillside. The mine was considered unique, perhaps owing its survival to the extremely poor quality of its copper ores and therefore to the lack of subsequent interest in mining them. When it was investigated in the 1960s, it was generally agreed that virtually all the evidence for early copper mining in Britain had been obliterated by later activity. Copper mining peaked in the late 19th century, by which time mechanised techniques were responsible for radical alteration of many mining landscapes. Survival of prehistoric evidence seemed unlikely. Today, this pessimism has been dispelled.

Thanks very largely to the (often unpaid) efforts of a small number of dedicated field workers, some 30 probable or definite prehistoric copper mining sites have now been identified in the British Isles, of which the Great Orme, with its visitors' centre and guided tours of the Bronze Age mine workings, is the most impressive. Many of these sites survive, despite all the odds, on surface outcrops of copper which, in the historic period, became well known and highly productive. In addition to Mount Gabriel and the Great Orme the best known and best investigated sites to date are at Parys Mountain in North Wales, Cwmystwyth in central Wales, Alderley Edge in Cheshire and Ross Island near Killarney in South-West Ireland.

Over the last decade or so, the antiquity of mining at these sites has been firmly established, mostly by radiocarbon measurements on charcoal and sometimes bone sealed within the mining waste. In addition to the mine, Ross Island also features a `work camp' area from which radiocarbon dates have been obtained. All of the sites were in use during the Bronze Age. Ross Island appears to be the earliest, with dates clustering in the second half of the 3rd millennium BC. This is just prior to the beginning of the Early Bronze Age in Ireland - the period associated with the introduction of metallurgy in the British Isles. The other sites were all in use at much the same time spanning the Early Bronze Age and earlier Middle Bronze Age from c 1900-1200BC. But what of the evidence for the copper they produced?

Bronze Age metalwork has an enduring fascination and has been the subject of study for two centuries or more. In the latter half of the current century, typological classification of metalwork has given way to a developing interest in its composition in the hope that stylistic or regional metal groups would share characteristic patterns of trace elements which might then be linked to particular ore sources. In the British Isles, significant changes in the impurity pattern of copper and bronze metalwork do occur over time and between different regions, but attempts to relate this to the general pattern of copper mineralisation in the British Isles have always been inconclusive.

Now, with the mines identified, it is becoming possible to develop a clear idea of the impurity patterns likely to have resulted from smelting the ores from particular places. A detailed mineralogical survey by Rob Ixer at the University of Birmingham is now revealing just such information. The work is painstaking and highly skilled. A detailed understanding of metallogenesis and ore geology are required even to select representative mineral samples for further study. Ore petrography is then used to identify the mineral suite and build up a picture of the formation process (or processes) and subsequent geological history of the deposit. Only with this level of understanding is it possible to identify the ore that was actually mined from a particular site.

The results of Ixer's analyses are fascinating. With one exception, all of the sites investigated can only have produced virtually pure (impurity free) copper. This contrasts strongly with the Bronze Age metalwork for which common impurity patterns have emerged.

The earliest metalwork, with a primary distribution in South-West Ireland, often contains considerable arsenic - sometimes several per cent - with lesser amounts of antimony and silver. Later, at roughly the time that mines such as the Great Orme and Cwmystwyth were in use, these compositions give way to copper with a higher nickel content. By this stage the copper is most often alloyed with tin to form bronze and has a wider distribution across upland areas of Britain.

Of the mines investigated, only Ross Island, the earliest, produced copper with a significant arsenic-antimony-silver impurity pattern. Could it be that Ross Island, perhaps together with as yet undiscovered mines in the region, was the dominant source of the earliest copper before it became mixed and diluted with the pure copper from Wales and northern England? Were Killarney's Beaker Culture people our first metallurgists? If so, where did the nickel come from in the later metal? Does it represent the growing status of alternative groups of metallurgists with their own as-yet-undiscovered copper supply?

Answers to some of these questions are now emerging from lead isotope analyses of the ores. The isotopic composition of lead within an ore deposit relates to its geological formation process and age, with the result that different deposits can have characteristic values (although they sometimes overlap). Lead isotopes are unchanged by the smelting process so that the signature of the ore is carried by the finished copper.

Brenda Rohl, working at Oxford University's Isotrace Laboratory, has analysed ores from many of the newly discovered mines as well as numerous Early and Middle Bronze Age copper and copper-alloy artefacts. Some of the earliest, `type A', metal tools do have isotopic signatures which match ores from Ross Island, but the mine is unusual in having two types of ore with quite distinctive isotope signatures. Some `type A' tools have isotope ratios which suggest they were made by mixing the two.

For the later metalwork analysed by Rohl the picture is more complex with a pattern indicative of the mixing of copper from multiple sources. Only at Ross Island is there evidence of prehistoric metal processing in Britain, and in general we do not know how far ores were transported for smelting. However, the mixing is less likely to have resulted from the original smelting process, and was probably rather the consequence of later melting-down and recycling of artefacts from different sources. This is a relatively simple operation and may have been performed, perhaps from an early date, more commonly than is traditionally thought.

Occasionally, however, very distinctive patterns emerge from which specific conclusions can be drawn. In one case analysis of five of the nickel-rich `type B' artefacts shows them to have a highly unusual lead isotope composition resulting from uranium associated with the ore. There are only a handful of deposits, all in Cornwall, where such nickel- and uranium-bearing copper ores occur.

The discovery of the copper mines has undoubtedly given a boost to archaeometallurgy in the British Isles, at last allowing us to bridge the gap between artefacts and their sources. Clearly much remains to be done, but interesting results are already emerging which reinforce the suggestion of an early metallurgical focus in South-West Ireland. Their products were soon joined by, and mixed with, those of other miners working the copper deposits of Wales, northern England and, almost certainly, Cornwall, where prehistoric mines may yet be awaiting discovery.

Dr Paul Budd is NERC Advanced Research Fellow in Science-based Archaeology at the University of Bradford

Locating the birthplace of St Patrick

Ireland's patron saint was born and raised in Somerset, argues Harry Jelley

Historians take great pleasure in locating the birthplaces of great figures from the past. This is particularly true when the birthplace has been the subject of heated debate for centuries, as is the case for St Patrick, the 5th century Romano-Briton who later became patron saint of Ireland.

There have been claims and counterclaims. Strathclyde was once thought a possible birthplace for Patrick, presumably because of its proximity to Ulster. Several sites in South Wales have been suggested, as well as the area near Birdoswald on Hadrian's Wall. For various reasons, however, none of these provides a satisfactory solution.

It is my view that Patrick was in fact born in south-western England, in Somerset, at or near the village of Banwell, five miles east of Weston-super-Mare. A settlement of late Roman date is known in the area. The archaeology of the region, which was highly Romanised, suggests Somerset was a plausible location for Patrick's family estate - unlike some of the remoter locations suggested - while placename evidence supports a Somerset location directly.

There is, in addition, an undated, unexplained earthwork in the form of a cross at Banwell, surrounded by a bank. A scheduled monument, it is listed by English Heritage as a Roman camp (which is unlikely), and is described in the local Sites and Monuments Record as a rabbit warren, for which there is also no evidence. I like to imagine the cross has a religious interpretation, constructed as a monument to Patrick by missionary Irish monks a few centuries after Patrick's lifetime, at a time when the memory of his birthplace survived.

The established facts of Patrick's life are few. He was captured aged 16 by Irish raiders and taken as a slave to Ireland, from where he is said to have escaped to Gaul before returning to Ireland as a missionary in AD432, founding Ireland's first Christian church at Armagh. Recent research disputes the traditional dates, suggesting Patrick may have lived a generation or so later.

In his Confessio, Patrick writes:

I had as my father the deacon Calpornius, son of the late Potitus, a priest, who belonged to the small town of Bannavem Taberniae; he had a small estate nearby, and it was there I was taken captive.

Further evidence of his childhood comes from his Letter to Coroticus, in which Patrick says his father was a decurion, a town councillor. The details are murky but the impression is of a well-to-do family, connected to a town, with a villa estate nearby.

`Bannavem Taberniae' is improbable Latin and it is usually amended to Bannaventa Berniae, a place that should - for reasons of historical plausibility - lie on the western side of Britain susceptible to Irish raids in an area of recognised villa settlement, such as Somerset.

Bannaventa Berniae is a Latinisation of a Celtic placename. Banna (root ban) is a Celtic word meaning bend, crook or peak. Venta occurs three times applied to tribal capitals - Winchester, Caerwent and Caistor in Norfolk (Venta Belgarum, Venta Silurum, Venta Icenorum) - where Venta indicates an area of local administration. Berniae arguably derives from a Celtic root bern meaning `gap' or `pass', as in the modern Irish bearna.

The further west one progresses in Britain the greater the survival of Celtic elements in modern placenames. Banwell, recorded in Asser's biography of King Alfred as Banuwille, and in a charter of AD904 as Banwylle, contains the Celtic element ban; and appropriately, overlooking Banwell, is an unusually shaped hill known as Crook Peak, perfectly fitting the meaning of the Celtic word. Wylle is Old English for `pool' or `spring'.

If Bannaventa Berniae became Banwell, we are assuming the retention of some elements of the original name, and the disappearance of others. This was a normal occurrence, and can be seen for example in the names of Winchester and Caerwent. However, around Banwell we do seem to find placename echoes of the lost original name. Venta is remembered in Winthill, where a villa was excavated in the 1960s. Winterstoke was a medieval hundred, of which Banwell was the capital manor. There is also a Winterhead hill, Wintreth in the Domesday Book, and the village of Winscombe (or Wintscombe).

As for Berniae, the 1841 Shipham Tithe Map portrays three fields named Bairn's Green, Bairn's Combe, and Little Bairn, all lying within a few hundred yards of the Winthill Villa. No trace of any family of this name can be found in the hundred years prior to 1841, and no earlier names of these fields are recorded. Can they preserve a memory of a Celtic placename element, bern? The idea is not impossible. Moreover, nearby lies the Churchill Gap, a pass through the Mendips used by the old main road from Bridgwater to Bristol (now the A38), which fits the meaning of the Celtic word.

During the 1960s two villas were excavated at Banwell. One, known as the Riverside villa, was a wealthy establishment with mosaics. The other, only half a mileaway, was at Winthill, with occupation lasting into the 5th century. The excavators found a post-Roman cemetery dug into the villa with bodies aligned east-west and without grave goods in the Christian fashion. Half a mile is unusually close for two villas, and evidence suggests that the Winthill site was not a villa, but a house in a small town. I believe this town was Bannaventa Berniae.

Fieldwork by local archaeologists over recent decades has produced large quantities of Roman pottery, coins, the odd burial, a Roman road surface, and building foundations around the `villa' site. For years, local opinion was that this was the site of an extensive Roman settlement, and this seemed to be confirmed when a pipeline was dug past the `villa' in 1994. The trench was overseen by AC Archaeology, who contracted local archaeologist Richard Broomhead to record and interpret anything revealed during the excavation. The narrow trench produced several substantial foundations of Romano-British date, part of a stone apsidal structure, and large amounts of pottery. The site has had no further excavation or geophysical survey.

A mysterious earthen cross lies in woods near Banwell. Its arms are about 20 metres long, aligned roughly on the four compass points, and the cross is surrounded by an almost square earthen bank. An excavation in 1961 to test its date and function proved inconclusive, but established that a continuous ditch originally lay on the inside of the bank, which appears to rule out the `official' English Heritage explanation of the site as a Roman camp.

The site is listed in the SMR as a rabbit warren. Earthen crosses, found mainly in Yorkshire, are usually interpreted as wind-breaks for livestock, and none has been shown by excavation to be a warren. The 1961 excavation showed that the underlying bedrock had been dug out to form the earthwork and make a ditch, which seems unnecessarily laborious for the building of a warren. Medieval windmills were sometimes built on cross-shaped mounds, but none as large as the cross at Banwell.

This configuration of cross and perimeter bank is in fact unique in Britain, and it seems not unreasonable to look for a unique explanation. Since Banwell lay within a Saxon royal estate which included the monastery given by King Alfred to Bishop Asser, and if Patrick had been born nearby, perhaps a religious interpretation is feasible.

Vigorous Irish evangelisation swept across Britain in the 6th-8th centuries, while at the same time the see of Armagh campaigned to establish its ecclesiastical supremacy in Ireland based on Patrick's supposed founding of Ireland's first church there. It is certain that there was an Irish interest in Patrick's birthplace, and in the work of the 7th century writer Muirchu we find: `Patrick came from the town of Bannavem Taberniae, not far from our sea . . . We have discovered for certain and beyond any doubt that this township is Ventre .. .'

I suggest that itinerant Irish ecclesiastics, or more likely residents of a monastery at Banwell or even at the Irish-inspired monastery at Malmesbury, erected a memorial at the place of Patrick's birth, in the absence of a known place of burial. Some six placenames around the cross containing the elements rod, rhod, or road could perpetuate the Old English rod, meaning `holy cross'.

A traditional Irish altar or grave, a leacht, a box shaped construction faced with stone, could have formed a cross if four were placed at right-angles, and this, I suggest, is what was created here.

There is an early tradition that the great abbey at nearby Glastonbury was jointly dedicated to St Patrick and the Virgin Mary. In the 12th century, William of Malmesbury mentions Patrick's `grave' at Glastonbury, and refers to the number of Irish pilgrims at the site. How can we explain this supposed Irish connection?

In 1091, the supposed remains of Saint Benignus were taken to Glastonbury from Meare in Somerset, as chronicled by John of Glastonbury. It may be that, in similar fashion, a dedication to Patrick at Banwell was taken over by Glastonbury, and later upgraded to a supposed `burial', a process completed after the great fire at the abbey of 1184 when a tomb to St Patrick was construced beside the altar.

Harry Jelley is an independent historian and author of Saint Patrick's Somerset Birthplace (Cary Valley Historical Publications, 1998, ISBN 0-9532501-0-5)