The following article was first published in the Spring 2021 newsletter of the North Devon Archaeological Society. The society’s newsletters may be read at the NDAS website.
I devote a lot of my spare time to drawing historical maps. My objective is to make maps that are comparable with the Ordnance Survey’s 1:25000 scale Explorer maps, but which show the landscape not as it is now but as it was about the year 1840, just before the railways arrived in this county. I am publishing this work as a series of maps which I call the Devon in 1840 series. So far, they cover northern Devon.
I mainly use the scale 1:16000, or about four inches to one mile. This is a large enough scale to show field boundaries and many field-names, while also making it possible to show quite a large area (covering several parishes in full) on one sheet of paper. In addition, I make larger-scale maps of the towns and some villages, and also of landscapes that have special historical interest such as Braunton Great Field and the parkland around Castle Hill House. These more detailed maps are printed on the other side of the sheet.
From an archaeological standpoint, perhaps the greatest interest of maps of the early nineteenth-century landscape is locating the many farmsteads and other minor settlements that were occupied at that time but later abandoned. However, in this article, I want to focus instead on features that went out of use long before 1840. My maps include what cartographers call antiquities, or features in the landscape of archaeological interest. I indicate antiquities by the time-honoured trick of labelling them with a distinctive typeface that evokes ‘oldness’. The illustration (Fig. 1) shows the earthwork known as The Beacon, high above the cliffs near Martinhoe, seen here in the context of early nineteenth-century fields. This earthwork is known from excavation by A. Fox and W. Ravenhill in the 1960s to have been a Roman outpost.
It is not obvious how to treat archaeology on a map of the landscape in 1840. How old should a feature be for it to be considered an antiquity? For instance, a fort built during the Napoleonic Wars might be labelled as an antiquity on a present-day map, but it would hardly have been treated that way in 1840, when it was not even forty years old. In fact, I have chosen 1651, the end of the Civil War, as a cut-off date: only features older than that may be marked as antiquities. This is one of several rules I have devised for deciding what antiquities to include on the maps.
One of the rules is that antiquities that were entirely destroyed before 1840 are not generally shown. In other words, a feature is shown only if there was something in the landscape that an observer at that time would have seen. Actually, I relax this rule in cases of exceptional interest. Some examples are on my town-map of Barnstaple, which is based on a map made by John Wood in 1843. Barnstaple was formerly a walled town with four gates, but only one of them, the West Gate, was still standing in 1843 (it was demolished in 1852). Nevertheless, I have added antiquity labels to mark the sites formerly occupied by the other three gates. The extract in Fig. 2 shows the site of North Gate.
As far as possible, I try to represent antiquities on the maps in the state they were in around 1840. Again, the guiding principle is that the map should show what an observer at the time would have seen. Of course, that may be very different from the appearance of the site today, and in fact the maps show many earthworks and stone monuments that have now been destroyed.
My challenge, then, has been to list antiquities that were extant in Devon around 1840, with enough information about each site to plot it on a map, appropriately labelled. Unfortunately, for the great majority of antiquities, no record was made until near the end of the nineteenth century, which is four or five decades after the period of interest. This brings us to the problem of sources of evidence. My maps are derived mainly from tithe maps, the large-scale parish maps produced mostly between 1839 and 1842. Tithe maps are an excellent source of information on buildings, roads, field boundaries, and parish boundaries, and the apportionment documents associated with them are a rich source for place-names and details of land use. However, tithe maps are usually a disappointing source for antiquities.
There are a few exceptions. The Braunton tithe map gives the exact position of St Anne’s Chapel, isolated among the sand dunes of Braunton Burrows. Described as a ruin by Lysons in 1822, the site is shown on the tithe map as a small rectangle, suggesting that some remains were still visible in 1840, although they disappeared soon afterwards. Another example is on the Lynton tithe map: in the moorland of Lyn Down, the surveyor marked the position of a monolith named Long Stone, and added a remark ‘nothing particular here but a stone about 7 feet high’. This must have been the standing stone still present near the site today (Fig. 3), although Lyn Down was enclosed around 1860 and the stone is now in a field and not at its original position. (It is not to be confused with the well-known Long Stone near Chapman Barrows.) But a few exceptions aside, tithe surveyors did not record earthworks or stone monuments on their maps because they were not relevant to the maps’ purpose.
The first methodical record of antiquities was made by Ordnance Survey with the publication of their County Series maps at scales 1:10560 (6 inches to 1 mile) and 1:2500. In north Devon, surveys for the first edition were made in the 1880s. The great majority of the plans of ancient earthworks on my maps have been copied from that source. By ancient earthworks, I mean those features that survive in the landscape only as mounds and ditches, including hillforts and settlement enclosures of Iron Age date, and castles of the motte and bailey type like Holwell Castle near Parracombe. Earthwork plans on County Series maps have sometimes been criticized as the surveyors were not archaeology specialists. For instance, the plan of the Roman fortlet at Martinhoe (Fig. 1) is notably different in shape from the plan made by the excavators in their 1966 report. However, County Series plans are sufficient to convey the size and general form of the monument, which is all that can really be shown on a map at 1:16000 scale. I also think that an earthwork that was intact in the 1880s would not have changed much since 1840, so it is reasonable to use the same plan on a map of the locality at that date.
County Series maps are also an essential source of information about stone monuments of various kinds, from Neolithic or Bronze Age standing stones to medieval roadside crosses. Those maps are a more problematic source for stone monuments than for earthworks, for two reasons. First, the authenticity of a stone as a prehistoric monument is sometimes questionable, and second, it is not unknown for stones to be moved, so that a monument shown at a particular place on the County Series map might not have been at the same spot in 1840. I have already mentioned the monolith on Lyn Down which is not now at its original position. To make a judgement about whether or not to mark an antiquity on the map, if its status is open to doubt, I have followed professional archaeological assessment of the site in the Historic Environment Record (HER) online databases. These comprehensive inventories, maintained by Devon County Council and by the National Park Authorities for Dartmoor and Exmoor, can be accessed through the Heritage Gateway website.
I should say that although I have made heavy use of County Series maps as source material for antiquities, I have not copied the descriptive labels on those maps, which were sometimes rooted in nineteenth-century amateur archaeology. For example, three sandstone blocks on Witheridge Moor would not be described today as ‘Druidical stones’, as they were on the County Series map of 1889. Ordnance Survey’s treatment of archaeology became more professional after 1920 with the appointment of an Archaeology Officer, although most large-scale map sheets for Devon were not revised until after the Second World War. Incidentally, because their status as prehistoric monoliths is doubtful, I decided against marking those stones on my Witheridge map.
Antiquities have always been at risk of loss, either by outright destruction such as the demolition of a building, or by erosion such as ploughing over an earthwork. No doubt some could be seen in 1840 that were then lost without any record. It follows that no map drawn today of the landscape at that time can give a complete picture of the antiquities that were actually present. There are tantalizing references to stone circles and the like in works by early topographical writers such as Risdon and Lysons (see L. V. Grinsell, The Archaeology of Exmoor, 37-49), but the state of those monuments in 1840, if indeed they had ever existed, cannot now be recovered.
Since the Second World War, archaeologists have discovered many sites of interest by searching aerial photographs for soil-marks and crop-marks that reveal buried structures. Often, nothing remarkable can be seen at these sites at ground level. Some sites of this kind might have been actual earthworks in 1840, but I have resisted the temptation to show them on my maps if there is no evidence that an earthwork was extant at that date. To give just one example, on Rowley Down, south of Parracombe, a soil-mark of a rectangular enclosure can be seen clearly on a number of air photographs taken in 1946 and 1947; but a map of Rowley Down made in 1797 does not show anything that corresponds to this enclosure, despite showing remains of some ‘old banks’ in a different part of the moor. Was the feature on the air photographs an upstanding earthwork in 1840? No certainty is possible, and I have generally followed the maxim ‘if in doubt, leave it out’.
Occasionally, there are discoveries of antiquities that are subtle but still visible features in the landscape today. Reports about them can be found by searching the HER online databases. I show these sites on the maps, if there is enough information to depict them with reasonable accuracy at the map’s scale. Some of them are not yet on Ordnance Survey maps. One example is a hillfort at Halsdon, to the west of Dolton, which was reported in 1980. Part of this earthwork is in a Devon Wildlife Trust nature reserve and I went to look at it in 2019. Located at the top of a ridge between the River Torridge and a stream, there is a bank and ditch in woodland on the south-west side, and the earthwork is traceable through pasture to the north where it is visible in air photographs. The east side has been lost to road improvement. All the information in Fig. 4 comes from the Dolton tithe map and apportionment, except the earthwork and contour lines. Notice that the field within the hillfort was named Berry Hill at the time. Field-names like Berry Close, Burrow Park and Castle Close often indicate some kind of ancient earthwork (possibly ploughed-out long before the tithe survey), and because of their significance I have made a point of including such names on the maps.