Ancient earthworks

Ancient Earthworks

Introduction

Excavation affords the only method of ascertaining the age of most earthworks. Not until the spade has disclosed the evidence which an ancient earthwork may hold in the relics which lie buried beneath its area or within its banks, can we as a rule assign its construction to a particular period. A striking illustration of this is afforded by the new light which was thrown upon the age of Hunsbury Camp, in Northamptonshire, by the unlooked-for discoveries during excavations there, as described in the volume for that county in this series. At the same time such thorough investigation has been possible in the case of only a very few of the large number of earthworks in the country, and some working method of classification is necessary in dealing with them; the one which is now finding general acceptance, which is adopted in the present work, is that set out by the committee formed by the Congress of Archaeological Societies, of which Mr. I. Chalkley Gould, F.S.A., is the honorary secretary. The classification is based upon the form of the work, and stands as follows:—

A.—Fortresses partly inaccessible, by reason of precipices, cliffs, or water, additionally defended by artificial works, usually known as promontory fortresses.

B.—Fortresses on hill-tops with artificial defences, following the natural line of the hill; or, though usually on high ground, less dependent on natural slopes for protection.

C.—Rectangular or other simple enclosures, including forts and towns of the Romano-British period.

D.—Forts consisting only of a mount with encircling ditch or fosse.

E.—Fortified mounts, either artificial or partly natural, with traces of an attached court or bailey, or of two or more such courts.

F.—Homestead moats, such as abound in some lowland districts, consisting of simple enclosures formed into artificial islands by water moats.

G.—Enclosures mostly rectangular, partaking of the form of F, but protected by stronger defensive works, ramparted and fossed, and in some instances provided with outworks.

H.—Ancient village sites protected by walls, ramparts, or fosses.

X.—Defensive works which fall under none of these headings.

Although this classification is one of form it to some extent provides a chronological guide which we may use tentatively, remembering that in most cases the period of the earthwork must be judged finally by the finds which may be made within its area. Even then it has to be remembered that fortresses were used by successive occupiers of the country, each of whom left their relics, so that a solitary find of any period proves little beyond the fact that it was used during that period, which may not be that of its construction. For instance, Roman relics are often met with on the hill-top forts, but the investigation of this class of work would lead us to believe that they are among the earliest of the whole series, although possibly enlarged and strengthened at later periods and temporarily occupied by Roman forces, whose peculiar fortification we have evidence for seeing in the regular rectangular camps found usually on low ground. Not until the earliest relic the work may hold has been extracted from it are we safe in assigning it a period. It is necessary to insist upon this on account of the unscientific way in which form has been made the test of age in the past, certain shapes being assigned to certain races with a finality which later evidence has shown to be quite unwarranted. And here, too, it may be well to warn the student of these works against the ideas of those antiquaries of the last generation who saw in them connected or opposing lines of fortifications covering a considerable tract of country, and arguing an organization which our knowledge at present hardly justifies us in assuming.

We are probably right in looking upon some of the promontory and hill-top fortresses as the earliest of the series, constructed in the Neolithic, or Later Stone, Age as places of refuge and defence for the inhabitants of the hut villages on the lower ground; while others may have been the work of a later period when bronze or iron had come into use; but here again caution is necessary, for we have learnt of late that we must be careful not to under-rate what the earliest inhabitants could achieve with their horn pick, bone shovel, and stone hammer. We must remember, too, that the Stone Age to which some of these forts are assigned was far longer than any succeeding age and probably covered thousands of years, and that the possibilities of achievement which had been reached by a long, if slow, development were of a very high order, as is instanced by the construction of Stonehenge. It is probable, then, that the earliest fort was the fortified hill-top, the shape of the hill dictating the line of the work, and that next in order chronologically but probably falling within the same protracted period were the forts on high ground, plateau forts as they are sometimes called, in which the line of the work is independent of the form of the ground, but which generally suggest by their construction and shape the hill-top fort. The use of the rectangular enclosure, though typical of the Roman camp, was not confined to the Roman period.

With the fortified mount and the mount and bailey fortress we are well within the historic age, but we cannot here discuss the question as to whether the Norman conquerors introduced it or found it here in a modified form. There is, however, no question that the vast majority of such strongholds are of post-Conquest date. The moated homesteads and enclosures appear so distinctly mediaeval that it may seem out of place to include them with ancient earthworks, but many of them may be survivals of enclosures round the holdings of the Anglo-Saxon settlers, and it is not improbable that some of them represent the later adaptation of the site of a much older work. I have ventured to call attention to characteristics of two in this county which would seem to support such a suggestion. The moats at Stratton Audley and Beckley may represent such an adaptation.

A glance at the map of Oxfordshire shows that the distribution of earthworks is very singular. The county is divided by the Cherwell Valley from Banbury to Oxford, and in the eastern portion of this division the works which we may consider prehistoric are represented solely by the Dike Hills at Dorchester, which constitute with the rivers there a fortress of the 'A' or promontory class, and the Roman station of Alchester. The station which is assumed to have existed on the site of the present village of Dorchester is not dealt with here, as the later village covers the site and has destroyed its works, and the relics which it may have left fall into another section of this work. There is also on this side of the county the earthwork in Wyfold Wood on the south, which we are unable to reconcile with either of the classes in the scheme given above. Nearly the whole of the works which we may consider the earliest are to be found, therefore, on the western side of the Cherwell Valley, and on this side the map shows that they are mostly confined to the north-west. With only two exceptions the works which we have put in class 'B,' the hill-top and plateau forts, lie north of a line drawn from Deddington to Shipton-under-Wychwood. This is understandable when the physical conditions of the county are considered. It is, of course, quite unnecessary here to insist upon the artificial character of the county boundary, an invention of much later date than the class of earthworks we are now considering, and with which consequently neither they nor the people who constructed them have any manner of relation. The northwestern boundary line of the later county has been carried over the high ground which constitutes the outlying spurs of the Cotswolds. Conical-shaped hills and plateau-like elevations, intersected by innumerable valleys, form the sort of country in which the hill-forts of early man are found, while the rich red loam of the northern portion is the richest of the county and makes it the best for habitation by a people who were already practising cultivation. The altitude above sea-level is 300 ft. at the Cherwell at Banbury, and rises to 750 ft. at the Rollrich Stones on the western extremity of the county, and crossing the border we ascend still higher as we approach the Cotswolds proper. These hills are covered with remains of prehistoric earthworks, and the Oxfordshire specimens—Madmarston, Tadmarton, Ilbury, Idbury, Lyneham, and Chastleton—must be considered as belonging to the Cotswold series.

Below Shipton-under-Wychwood we reach ground of lesser altitudes, and here a great belt of forest stretched right across the county in ancient times, comprising the forests of Wychwood, Woodstock, Shotover, and Bernwood, the latter in the north-east about Bicester and connecting with Northamptonshire forests. South of this ran the low plains of the Oxford clay along the Thames Valley from Bampton to the inhospitable and dreary waste of Otmoor on the east, and the lowlands give place on the south to the wooded heights of the Chilterns. It must not, however, be understood that the north-west was the only district inhabited by prehistoric man. A reference to another section of this volume will show that his relics have been found in all parts of the county, while the barrows and tumuli which represent his burial mounds are found scattered all over the portion west of the Cherwell. Of the forty-five existing specimens only seven are found in the half of the county east of this river. This, coupled with the fact that all the megalithic remains are in the western half, and that other remains of this age appear more numerous there, would seem to suggest that the broad Cherwell Valley formed the natural eastern boundary of a populous tribe occupying the district of the Cotswolds. So far, however, as the absence of the barrows in the eastern half is concerned, it is necessary to remember that it has been much longer under the plough than a large portion of the other, where the district between Burford and Charlbury covered by the Wychwood Forest and in which a considerable number of the barrows are found was only disafforested in 1862, and that in a very few years many of the barrows in the part now brought under cultivation will have ceased to exist.

The rectangular earthworks, except those in the parishes of North Newington and Chadlington in the north, are extremely small and almost give one the impression of miniature works connected with a Roman settlement on the Akeman Street and within the area encompassed by the dyke called Grim's Ditch, which exists to the north of this road in the western half of the county, in which area Roman remains are plentiful.

In many parts of the country are to be found remains of lines of entrenchment several miles in length and generally called dykes. There are traces of three in Oxfordshire, one referred to above; another, also called Grim's Ditch, in the extreme south of the county; and a third, called by the three different names of Avesditch, Ashbank, and Wattlebank, running from where the Akeman Street crosses the Cherwell to the northern boundary of the county at Souldern. As with the camps, so with the dykes, excavation must provide the final word as to their age. In the case of the Wans Dyke, an extensive entrenchment in the south-west of England, the theories of the antiquaries, which made it a Celtic work, were upset by the investigations of General Pitt-Rivers, which showed it to be of Roman or post-Roman construction. While therefore refraining from any explicit pronouncements respecting the Oxfordshire dykes, it may at the same time be helpful if we suggest an interpretation of their characteristics which may provide a working basis for their future investigation. Two of them, the Grim's Ditch, which we call 'A' to distinguish it from the work bearing the same name in the south, and the Avesditch, seem to have a connexion with the Roman road, the Akeman Street, which enters the county from Aylesbury to the south-east of Bicester and leaves it below Burford. The Avesditch commences at the ford of the road over the Cherwell, and runs north on the high ground to the east of the river, while the Grim's Ditch describes a semi-circle to the north with the road as its base. A study of the banks of these dykes shows that the Avesditch would appear to have faced west, and the Grim's Ditch north. We have shown above that the half of the county west of the Cherwell would appear to have been an area of greater population in pre-Roman times than that to the east, and a conjecture regarding the dykes, which is feasible pending further investigation, is that they were made by Romans who came from the east and constructed the Akeman Street, and were made against the Britons of the north-west area of the county. An examination of the Grim's Ditch and the Roman remains, which its sweeping line covers, led General Pitt-Rivers to assign it to those invaders. To the south of Akeman Street, and at almost an equal distance from it as Grim's Ditch is on the north, is the small line of entrenchment at North Leigh, which forms the southern boundary of the area of Roman remains which has its centre at Stonesfield. This entrenchment faces south, and three-quarters of a mile to the south of it is the ancient camp in Eynsham Park, and it would appear, therefore, to have been part of the scheme of protection by the makers of the road, and to have been part of their defence against the British population, of less numerical strength perhaps than that to the north, which occupied the country between the forest and the Thames.

Turning to the southern Grim's Ditch, which we distinguish by labelling it 'B,' we find the conditions analogous to those described in connexion with the northern work. The Icknield Way also runs from near Aylesbury, enters Oxfordshire at Chinnor and skirting the base of the Chilterns leaves it at Goring, and across its course from Wallingford to Henley ran the Grim's Ditch, cutting off the hilly country encompassed by the southern dip of the Thames between these two towns. Its banks show it to have faced south, and although the Icknield Way has the characteristics of an ancient trackway rather than a Roman road, it is a suggestion equally feasible as that thrown out in connexion with the northern works, that it may have been used by a Roman invasion, starting from the same point near Aylesbury as that which laid the Akeman Street, and which may have thrown the Grim's Ditch across the Chilterns to separate itself from the British population of a portion of their hills and woods. In the centre of the area so cut off by the Ditch is the earthwork in Wyfold Wood, and although we have been unable satisfactorily to include it in any class of the scheme of classification adopted, it may be ancient and have been a fortress of the people against whom the Ditch was constructed.

The whole of the county came in time under Roman domination, and the remains of that people are scattered all over its area, but it is rather significant that the only two sites which can be looked upon as more than those of villas are both in the eastern half, at Dorchester, a little to the north of the commencement of Grim's Ditch, on the south; and at Alchester, in the rear of the Avesditch, on the north. These two were connected by a Roman road. The evidence of the earthworks and dykes, as we are at present able to read it, would appear to be that the sparsely populated plains on the east were first occupied by Roman invaders coming from the Vale of Aylesbury, and operating afterwards against the British population of possibly greater account in the hills and woods of the west and extreme south. The two local entrenchments on Gravenhill and Swyncombe Downs appear to have been constructed against forces advancing along the lines of the Akeman Street and the Icknield Way respectively. At the same time, it must not be forgotten that the county suffered several other invasions by later forces, and the above hypothesis is after all of only a tentative character.

Promontory Fortresses

(Class A)

Dorchester-on-Thames: Dike Hills.—At Dorchester the River Thames, after running due south, takes a turn at right-angles south-eastwards, and 1,070 yds. further on receives the River Thame, which runs into it from the north. The flat meadow and arable land south of the village of Dorchester is thus enclosed on three sides by the rivers; and across the north end of this space, just south of the village, runs a double row of banks, from the Thames on the west to the Thame on the east, and parallel to the southern line of the former river, thus enclosing an area roughly rectangular in shape of about 114 acres. These banks are known as the Dike Hills. They have unfortunately been mutilated, and the plough has reduced a portion of them, but they still exist, or may be traced, throughout the whole line of their original course. The most perfect part, which seems to have escaped any destroying agency, is that marked X—Y on the plan. Here, as shown in the section A-B, the northern bank is the higher, and this was apparently so throughout. There are traces of a shallow ditch on the north of this, 6 ft. wide and 2 ft. below the adjoining field. The height of the bank is 16 ft: 6 ins. from the bottom of this northern ditch, and 15 ft. 6 ins. from the level of the ground between the two lines of banks. The banks are 55 ft. apart, and the height of the southern one is 13 ft. on the inside, and 8 ft. 6 ins. from the level of the ground enclosed by the dikes and the rivers. This section of the banks is still covered with turf, but has been cut through at its eastern end, which is nearest the village. The northern bank apparently ends here in three mounds, which are due probably to its disturbance in modern times, but it is to be traced again for about 70 yds. within the corner of the hedge of the adjoining field—in, however, a very reduced condition. The southern bank has also been cut away at this (the eastern) end, but its course, in the shape of a low bank, sloping northeast, is discernible, and it appears to have run more to the south, pointing for a bend in the River Thame, at a distance of about 50 yds. from which it makes a return south-westward. This interval between the southern dike and the river, which appears to have been constructive and not the work of destruction, suggests whether the entrance to the enclosure may not have been at this point. A careful study of the ground, which is meadow-land, does not afford sufficient evidence on which to lay down the likely course of the northern bank at this end in its nearer approach to the river. Westward of X on the plan is, unfortunately, arable, and the plough is taken over the reduced banks and is gradually completing its work of destruction. They are still, however, very plain and, indeed, are striking features in this perfectly flat piece of land. Two gaps have been cut through to carry cart roads at W and X. Between these points the height of the northern bank has been reduced to 4 ft. above the field on the north, and 6 ft. 9 ins. on its southern side; while the southern bank is only 5 ft. high on its northern face, and 2 ft. 6 ins. above the field on the south. The tops of the banks have been much flattened, and while at section A-B the tops of the banks are only 3 ft. or 4 ft. wide, here the width of the northern is 15 ft. and the southern 29 ft., while the flat intervening space has been filled up to only 28 ft. wide. This middle section would seem to have suffered most, although we have no evidence as to whether the banks were originally of uniform height throughout their course. Westward of W, towards the Thames, the heights of the northern bank are 3 ft. 6 ins. and 7 ft. 9 ins., and the southern 7 ft. and 6 ft. on their north and south faces respectively. Here again the plough has done its usual work of flattening the tops and filling up the intervening space, the summits of the two banks, north and south, being 16 ft. and 27 ft. wide respectively, while the width of the space between them is only 24 ft. The Dikes end at a hedge 100 ft. from the present bank of the river. In 1870 the farmer employed his labourers in reducing this end of the banks. (fn. 1)

In the banks where they have been subjected to agricultural operations human remains and British, Roman, and Saxon coins have been found. (fn. 2) The River Thames in that part which encircles the area has also yielded a British shield and spear at Day's Lock; (fn. 3) two bronze spear-heads, while making a new lock in 1871; (fn. 4) human bones in the Oxfordshire bank, 6 ft. below the surface, south of the eyot at Day's Lock in 1864; (fn. 4) and a Celtic buckler and bronze dagger in the river at the site of the ancient ford at the junction of the two rivers in 1837. (fn. 4) A bronze Saxon buckle of a sword-belt, a spindlewhorl and other objects were found in a mound at the south-east end of the Dikes in 1874, (fn. 4) and General Pitt-Rivers found a quartzite pebble rubbed to an obtuse edge at one end and trimmed flakes within the earthworks, and a flint arrow-head has also been found. (fn. 5)

Within the area enclosed by the Dike Hills and the rivers Mr. F. Haverfield has noticed in the crops marks of circles and rectangles which in other places in the vicinity have proved to be the sites of pre-Roman and Roman settlements. (fn. 6)

On the Berkshire side of the Thames, opposite to the Dike Hills, are the Sinodun Hills, the summit of one of which is crowned by a hill fort.

Fortresses on hill-tops following the line of the hill

(Class B)

Bladon.—In the woods on Bladon Heath are the remains of an earthwork, called locally 'Round Castle,' which would seem to fall into this class although the hill on which it is situated is low. Round its summit, but now difficult to trace owing to the plantation which covers it, are the remains of two lines of entrenchments. Of the higher only a portion remains, but this is better preserved than the second line below it, which is slight. The average height of the bank of the upper line is 5 ft. from the top to the bottom of the fosse.

Deddington: Ilbury.—A quarter of a mile due south of the hamlet of Hempton is Ilbury Hill, a pear-shaped formation, the narrow portion pointing north-west. Along the foot of the western side runs a stream, and the slope of the hill is much steeper on that side than on the other. The 400-ft. contour line runs round the hill just below the summit. The long axis, which lies north-west and south-east, is 1,050 ft. long and the width of the top of the hill at its broadest part is 500 ft. A single line of scarping runs round the hill following its shape. The height of the bank varies, and has been much reduced on the eastern side by ploughing, but is distinctly traceable all the way round. It is best preserved on the western side where the ground is not cultivated and a hedge has been planted along it. At the northwest point (section A-B) the bank is 13 ft. high on the outside and 1 ft. 6 ins. on the inside. There is here a slight trace of a ditch 13 ft. wide, the outer bank rising 1 ft. 6 ins and then falling with the natural slope of the hill. The entrance is on the western side, and from it there appears to have been a way made down the slope of the hill to the brook.

Swalcliffe: Madmarston.—About 300 yds. due north of Swalcliffe mill a truncated conical hill rises from low-lying meadows through which runs the stream which works the mill, surrounded by other hills of similar height. The top of the hill is a flat plateau, the axis east and west being 550 ft., and that north and south at the longest part 450 ft. Round the summit run lines of entrenchments, apparently originally three in number. The top of the hill is ploughed, and this no doubt has destroyed the vallum round the summit, supposing one to have originally existed. The planting of a hedge round the top of the second line on the east side has obscured the fosse there. The three lines are traceable on the south and west, but the lowest ceases at a hedge at the north-west corner. The other two lines can still be traced all round. The most perfect remains of the triple work are at the south-west. The uppermost ditch there is 8 ft. 6 ins. below the summit; the second bank rises from it and drops 4 ft. 6 ins. into the second ditch. The entrance is on the south, and here the works have been carried down the slope of the hill and turned inwards, the lowest of the lines here being on the 500-ft. contour line. The works of the entrance still exist, although a plantation which now covers this part of the hill makes it difficult to trace them. About a quarter of a mile due south of the camp and in a line with its entrance on rising ground on the opposite side of the stream formerly stood a round barrow, now destroyed, and also on that side of the stream three stone cists have been disinterred, each containing a skeleton at full length, face downwards. (fn. 7) The barrow was opened before it was destroyed and was found to contain ashes and burnt fragments of wood, but had apparently been previously disturbed. (fn. 7)

Fortresses on high ground but not dependent on natural slopes

(Class B—continued)

Chastleton.—On Chastleton Hill, which forms the south-western end of the long high ridge on which stand the Rollrich Stones, which are three miles to the north-east, is a roughly circular camp called locally 'Chastleton Barrow.' Along the summit of this ridge is an old trackway—in this part the road to Stow-on-the-Wold—which in the other direction runs past the stones, through Tadmarton Camp and by Banbury to Northampton, as described under Tadmarton Camp. About a quarter of a mile to the west of this trackway and in about the middle of the ground forming Chastleton Hill at an altitude of 700 ft. above sea-level is situated the camp. It consists of a single rampart of the uniform height of 9 ft. and 13 ft. on the interior and exterior respectively, covered with turf and planted with trees. The enclosed area is ploughed. The diameter of the rough circle described by the rampart is 450 ft. A bridle road runs through it, probably by its original entrances as there are no other openings in the rampart. Excavations were made here by Messrs. E. W. Brabrook, A. White, and J. E. Price, and described by the last-named in the Journal of the Anthropological Institute, 1881. The report states that the ramparts are formed of blocks of oolite, the natural stone of the district. In making sections of the ramparts deposits of pottery, burnt bones (all so broken as to suggest that they were used for the purpose of food) and charcoal, a bone awl made of the tibia of a small animal, a flint flake, burnt pebbles and various burnt shells were discovered. The explorers' description of the camp as rectangular appears to be an error, and their suggestion that it is of Roman origin seems to lack evidence. (fn. 8)

Eynsham.—In Eynsham Hall Park is an interesting earthwork of somewhat unusual character. In Woodley's Copse, a small wooded enclosure in the centre of the park, at the head of a shallow natural defile leading from the south and represented on the Ordnance Survey Map by the 300-ft. contour line, is an irregular-shaped earthwork consisting of a single rampart which rises on an average 3 ft. from the level of the enclosed area and falls 9 ft. into the ditch, which is of depth varying from 2 ft. to 4 ft. below the level of the outside ground. The entrance was at the south-east corner. The south-west and south-east sides conform to the contour of the ground, which slopes into the valley. The ground to the north of the camp is for some distance practically level, while the summit of the rising ground on each side of the valley at the end of which the camp stands is somewhat higher than the work itself. Looking at this part of the earthwork alone it would appear to be a defensive work against any approach from the south under cover of the small valley. From this, however, runs a second line of embankment which has the appearance of an addition to the original work. This bank is only 3 ft. high throughout its course, the ditch on the exterior is very slight and all round its course it has higher ground without than within. It leaves the earthwork first described at the south-east corner so as to include the entrance and turns to the south and then south-west, following the lowest ground of the valley until this turns a little eastward. At this point the planting of a spinney has destroyed the bank, but here it turns at right-angles to the north-west and ascends the rising ground until it reaches the level of the interior of the smaller work, towards which it then makes another turn. It does not, however, quite join it, and it may be that this gap represents the entrance. The slight character of this bank and its position with regard to the ground levels make it unlikely that it was part of a defensive scheme. It has more the appearance of an enclosure for cattle, but its situation in relation to the camp is in this case curious, for it is on that side of it on which apparently an attack was to have been met. It is, too, dominated by higher ground, while on the other sides of the camp is high ground on which such a protective enclosure could have been made. The suggestion which would meet with least objection would be that it was made as an addition to the original work as a place for securing cattle after the need for the latter as a place to meet an attack from this particular direction had passed away. A small pond just below the entrance to the camp fed by a stream from the north suggests the original water supply, but, if so, it is curious that it has been left outside the second enclosure, but the pond itself may be of later construction and the stream may originally have been led within the bank. The arrows on the plan indicate by their direction the fall of the ground and will help to make the situation clearer. About three-quarters of a mile to the north-east was a small circular camp now destroyed, and about the same distance to the north-west is a small line of entrenchment, both of which are referred to under the headings for those classes.

Idbury.—Half a mile to the south-west of the village of Idbury and on the north of the road from the village at its junction with that from Stowon-the-Wold is a roughly circular camp, the diameter of which is 700 ft. It lies on the edge of a tongue of high ground which falls on all sides save the north and northwest. The altitude is given on the Ordnance Map as 646 ft. The ground is under the plough and the work is very much reduced thereby. It consisted apparently of a single bank, with the entrance on the north-west.

Lyneham.—In the parish of Lyneham but situated by the side of the Chipping Norton road, one and a quarter miles north-east of the village, is a roughly oval camp consisting of a single rampart 9 ft. and 6 ft. high on its exterior and interior respectively, without a fosse. It is situated on the edge of a hill which falls from an altitude of 679 ft. above sea-level to 317 ft. at the River Evenlode, two miles distant. The hill forms a spur from the plateau and falls on all sides save the north and overlooks miles of country. A quarry has been cut into the rampart on the south, and the sections which have there been disclosed show that it is made of earth, although the soil of the district is so full of rubble that it has the appearance rather of having been constructed of small stones. In 1842 several human skeletons were found in this quarry, and human bones were also found close to the camp in 1875. (fn. 9) Two Saxon spear-heads, similar to the one found in the long barrow which stands on the edge of the hill 350 yds. to the south-west were found in 1884, when some trees were planted at the north-east corner of the camp. (fn. 10) The entrance is on the north side. Locally the work is known as 'The Roundabout.'

Tadmarton.—This camp is situated on Tadmarton Heath, one and a quarter miles due south of the village and two miles due south of the camp on Madmarston Hill. It stands on the highest part of Tadmarton Heath, but its shape is independent of the ground, which is level for about a quarter of a mile to the west, north, and east, and then falls gradually. The Ordnance Survey Map gives an altitude of 641 ft. for the centre of the camp. In shape it is roughly circular, the eastern side, where is the entrance, being compressed. The diameter taken south-west and north-east is 550 ft. There is a double line of vallum. The inner is well preserved throughout, but the outer disappears on the eastern side near the entrance. A road cuts through the camp, and its area on the northwest side of this has been planted as a fox-covert, but the works are still distinct, excepting on the west, where the outer vallum has been destroyed by the planting of a modern hedge. On this side of the road the height of the inner vallum is 4 ft. above the area of the camp and 7 ft. above the ditch, the width of which varies, but averages about 40 ft. from bank to bank. The outer vallum rises 2 ft. from the ditch on this side of the work, but is 37 ft. wide, with a flat top, and has a second ditch of similar width to the inner. The outward bank of this second ditch is 3 ft. high, and beyond it appears to be level ground, which is so thickly grown with bracken, however, that it is difficult to tell whether there was originally a third vallum. The area of the camp on the south-east of the road is still open heath, and the works here are easier to trace. The inner vallum rises on an average 3 ft. from the surface of the area and drops 5 ft. into the ditch, which on the south side is 50 ft. wide from bank to bank. The outer vallum is 3 ft. high. This is much narrower than it is on the north side, and although it appears to expand in width on the south this is probably due to a modern cart-track to the neighbouring farm being run up and along it at this point. There is no trace of an outer ditch on the south and east sides. There is a break in the inner vallum on the north side suggesting a second entrance, but this may be modern.

At the south-east of the camp and at a distance of 150 ft. from the outer vallum there is a small work of irregular form. The western side, running north and south, is 150 ft.; its northern side is 100 ft.; its eastern, 200 ft.; and its southern, 150 ft. It was apparently surrounded by a ditch which still remains in a slight form on the west and north. It is a much slighter work than the camp and forms a small platform 3 ft. above the ground level on the south and east and the bottom of the ditch on the north and west, being one foot higher than the surrounding ground on the two latter sides. There is no trace of a bank and it was probably a small stockaded enclosure.

The road which traverses the camp is an old trackway which runs from Northampton, by Hunsbury Camp and other earthworks in that county, crosses the Cherwell at Banbury, passes over Tadmarton Heath, turning due west about a quarter of a mile past the camp and, running close to a smaller camp now destroyed, continues past the Rollrich Stones, where it runs between the circle and the outlying monolith, passes close to Chastleton camp and joins the Fosse Way to Cirencester. The fact that it cuts through the camp on Tadmarton Heath may be taken as evidence that the latter existed at the making of the road.

Rectangular or other simple enclosures, including forts and towns of the Romano-British period

(Class C)

Chadlington: Knollbury.—This is a rectangular earthwork on the north side of the road from Churchill to Chadlington and about one mile north-west of the latter village. It is on the side of a hill sloping south-east and consists of a single bank with no trace of a ditch. Its form is that of a parallelogram, the major axis, which lies northwest to south-east, being 500 ft. and the minor 300 ft. It is characteristic of the rectangular earthworks in this county that they are laid out so that their sides run north-west to south-east and south-west to north-east, as in this case. The banks here are covered with turf, although the enclosed area and the field are ploughed, and they are well preserved excepting that on the south-east, of which only two fragments remain. The entrance was probably on this side as there is no trace of an opening in any other. The north-east side is bowed outwards at about half its course. The Ordnance Survey 551-ft. mark is at the north corner. The ground within the area is level and is rather higher than that without. An idea that the banks are built of stones is probably due to the stony nature of the soil which is full of rubble, as at Lyneham Camp.

Cornbury Park.—On the eastern side of the park on the high ground above the River Evenlode are the remains of three small earthworks (see map of course of Grim's Ditch 'A'). The first is at the north-east corner of the park, close to the wall nearest to the town of Charlbury. Parts of three rectangular sides remain. The complete side runs north-west to south-east and is 270 ft. in length, consisting of a small bank, 3 ft. 3 in. high on the exterior and 2 ft. on the interior, with a ditch which is 2 ft. below the outside ground. At the north and south ends the bank turns at right angles to the north-east and is continued on the north for about 50 ft., and on the south for 200 ft. On the north and east the ground falls rapidly. At about 180 yds. south-east of this there is a small tumulus. About half a mile south of this work are the remains of another, similar in structure to the one just described, and with its sides conforming to the same compass directions. In this case the south-east side is complete and is 100 ft. in length, the bank being 3 ft. high both on the exterior and interior. The entrance was apparently in the centre of this. It has a ditch on the exterior. This bank turns at right angles north-west at each end for a short distance only, both ending abruptly without any apparent reason. On the east the ground falls rapidly to the river and on the opposite side of the valley on the bank above the railway the Grim's Ditch comes to an apparent end pointing straight for the broken ends of the earthwork. At a distance of 250 yds. south of this work is another and a larger tumulus. The third earthwork is 130 yds. south-west of this tumulus and consists of a ditch 3 ft. 6 in. deep running north-west to south-east for 250 ft. At the south end it turns north-east at rather an obtuse angle and can be traced for about 100 ft. (see map of Grim's Ditch 'A').

North Newington: Castle Bank.—About three-quarters of a mile north-west of the village and nearer Wroxton is a simple rectangular enclosure called 'Castle Bank.' It is 570 yds. to the west of the road from Wroxton to North Newington at the point where French's Buildings stand. It is on level ground but on the edge of a small ravine or 'bottom.' It consists of a single bank which is well preserved on the north-west side, where its height on the exterior is 9 ft. Along the top of this bank runs the hedge of the field in which the remainder of the earthwork is situated, and as this is arable the work is being rapidly reduced and even where it is most distinct the bank is only 2 ft. 6 ins. high on the exterior and 1 ft. 6 ins. on the inside. The south-west side, which is slightly bowed outwards, is 500 ft. long, the remaining sides are 450 ft. each, that on the north-west being slightly bowed inwards. On this side there are slight traces which suggest that an outer ditch may have existed, but a footpath runs along its site on this side and if so has probably obliterated it, while the plough has destroyed it on the remaining sides. On this side the ground falls steeply to a small stream called Padsdon Springs. The entrance was in the middle of this side, nearest, as is often found to be the case, to the water supply. The 500-ft. contour line runs along this side of the work.

Piddington: Muswell Hill.—On the summit of Muswell Hill, which rises 392 ft. above the level of the village of Piddington, the Ordnance Survey altitude at the top being 649 ft., and which commands a most extensive view on all sides but the south, where it is intercepted by Brill Hill, is a rectangular enclosure named 'The Wilderness' and popularly called a Roman camp. It consists of a single bank forming a square, the sides of which are 250 ft. in length and the height of which is 2 ft. 6 ins. above the interior and 4 ft. 9 ins. on the exterior. The sides run due north to south and east to west. The whole space is turf-grown and the banks are admirably preserved and have certainly not the appearance of antiquity. It is difficult to look upon this as a defensive structure. The hill juts out into the plain northwards in a commanding promontory, but it is not on this that the work is placed. Neither is it on the highest part of the hill, the 600-ft. contour line running along the eastern side, and the higher ground on the south slopes down to a level with the top of the bank on that side, which has, therefore, no exterior slope at this point, and the interior of the area is absolutely dominated from the outside.

Spelsbury.—There is a small square earthwork in this parish although three miles from the village to the south-east. It is two miles due east of the town of Charlbury and 270 yds. north of the road from that place to Woodstock, in a field south-east of Ash Copse. Its sides are only 130 ft. in length, and there is a single bank which has a fall of 3 ft. 6 ins. into a shallow ditch which can be traced on two sides and part of a third. The south-west side has been dug into, apparently for quarrying. Like other rectangular works in this county its sides run north-west to south-east and south-west to north-east. It is within the area of the Grim's Ditch and is to be compared with the small works in Cornbury Park, which would seem to have some connexion with the Ditch. It may further be noted that both this work and those in Cornbury Park are the same distance (just under two miles) north of the Roman road, the Akeman Street.

Wendlebury: Alchester.—The most extensive Roman site at present traceable in this county is that of the station Alchester, situated in the parish of Wendlebury, one and a quarter miles south of Bicester, opposite the point where the road from that town to Oxford makes a sharp turn to the west. The road to this point is the Roman Way, which continued straight on through the centre of the station and can still be traced to Dorchester-on-Thames. The greater part of the site is, unfortunately, cultivated, but the course of the raised ways through it can still be traced by the low banks which vary in height above the ground from 15 ins. in some places to 3 ft. in others. In that part of the field where they have been destroyed, shown by the dotted lines on the plan, their course can still be laid down, as it is possible to note where they commenced by the breaks in the other banks, and the direction of the furrows has been changed to run parallel with them. At the south-east corner there is a mound 4 ft. 9 ins. high, and at the north-east corner there are the remains of another over which the hedge of the field has been carried. In the meadow to the west of the site, 80 yds. from it, is a roughly circular mound 200 ft. in diameter, and apparently rising from within a square base, called 'The Castle.' Excavations have shown that this covers the remains of a Roman building. We are only concerned here with the remains of the station falling under the classification of Earthworks, and it must be sufficient to state that Roman remains in the shape of foundations, coins, pottery, and other articles have been found all over the site, together with many human skeletons.

Fortified mounts with traces of one or more attached Courts or Baileys

(Class E)

I have found no examples of class D, that is of the simple mount with an encircling ditch or fosse, in this county. All the fortified mounts which now exist either have or have had one or more courts attached, and therefore fall under class E. In three of the examples given the type is not very pronounced, and it is possible that we see it in a stage of transition. At the same time it has appeared to be sufficiently traceable to warrant their inclusion in this class. At Chipping Norton the mount is of such a size and shape as to present the appearance rather of a raised court than a simple mount. At Deddington it presents the appearance of a mount from the eastern sides only, its height on the western sides being very little above that of the adjoining court, while at Mixbury it has practically been reduced to a corner tower in the rampart.

Ascot-under-Wychwood.—In the grounds of the Manor House is a circular mound, the summit of which is 42 ft. across in the direction shown in the section on the accompanying plan, and 46 ft. along the diameter at right angles to this. The average height is 8 ft., being highest (10 ft.) on the east, where a broad ditch separates it from a broad raised court, and lowest (5 ft. 6 ins.) on the west, where the ground has apparently been made up at or since the erection of the adjoining buildings. The bank and ditch of a large court remain in the adjacent field on the north-west side. The return at the east corner near the mound suggests that it did not include it, and the farm buildings and yards have destroyed all traces of the return of the ditch to the mount on the south. There was apparently, however, from the remains of a shallow bank between the ditch and the railway on this side, a second ditch which would appear to have joined that to the east of the mount and so have enclosed the whole work. Its further course has, however, been destroyed by the gardens of the house.

Banbury: Castle Gardens.—Banbury Castle, which dated from the Norman period, was totally demolished after the Civil War in the seventeenth century, and no plan, drawing, or detailed description made before its destruction is known. Its site is now occupied by gardens and by streets of houses, but a careful study of the ground with the aid of a plan of the property made in 1685, after the ground had apparently been cleared of buildings, (fn. 11) leads to the conclusion that we are dealing with a fortress which was in its origin of the mount or mount and bailey type. The making of streets and the cutting of the canal through the site have destroyed much of the evidence of the ground, but it is possible, from traces found in digging for draining and building, to lay down the course of the outer moat. It appears to have had four roughly equal rectangular sides enclosing a space of about seven acres. In the centre of this area is a mound 9 ft. above the ground-level at its base, now occupied by a rather ruinous cottage built at the demolition of the Castle on to the only remaining portion of its wall. This mound, according to the plan above referred to, was surrounded by a ditch now filled in, but the course of which was apparently immediately round the base of the mound and enclosing an area of 3 roods 3 poles. At the south-west corner the mound is now disguised by the streets which run up to it, but the ground here has been partly made up in recent times, while the road from the Market Place was the original roadway into the Castle. On the other sides its form is quite distinct. The mound is probably mainly natural, but increased or emphasized artificially. The regularity of the outer moat, which is square and runs at equal distances on all sides from the inner work, does not correspond with the usual character of the baileys in this class of fortress, and leads to the suggestion whether it does not represent a later extension of the original Castle, which may have been of the simple mount type, or have had a small bailey attached which was destroyed in the later work.

Chipping Norton: The Castle.—This consists of a large and high mound 500 ft. long by 250 ft. in the widest part, standing from 14 ft. to 20 ft. above the surrounding fosse, and divided by a small bank, which runs threequarters of its width across, into two unequal courts. Excepting a small piece at the north-east end there is no trace of a raised bank round the edge of the mound, the ground on the top of which is very uneven, as though it covered the remains of destroyed buildings. On the ground on the east below the mound another court is enclosed by a bank which rises 6 ft. on the interior and joins the fosse round the mound at the north-east end, where its exterior altitude is 14 ft. At this point there is a third bank which rises 1 ft. 3 ins. from the second ditch, and falls 6 ft. 6 ins. to the ground outside. At this point the summit of the castle mound is, therefore, 27 ft. 3 ins. above the exterior ground. The ground at this point and on the north-west slopes to the level of the stream. At the south-west end it also falls steeply to the stream-level, and here 20 ft. below the summit of the mound is what appears to be a small natural spur of the hill, on which a modern residence has now been built, but which originally may have been incorporated in the work. Below the foot of this and the stream are low banks, evidently the remains of a pool. The altitude given by the Ordnance Survey at the stream near these is 500 ft.; on a level with the church south of the Castle it is 597 ft., and a quarter of a mile further east it has risen to 710 ft. The work, therefore, was constructed on the slope of a steep hill, and would be entirely commanded from above.

Deddington.—Here there is an extensive series of banks covering an area of about sixteen acres. In this case the mound, marked on the Ordnance Survey Map 'The Keep,' consists of an irregular shaped court but little higher than the level of the ground in the interior of the court on the west, but rising on its eastern side to a bank 6 ft. above the level of its area and then dropping 23 ft. into the fosse. The general section X-Y on the plan will make its construction clearer. The court on the west is surrounded by a steep bank which on an average is 6 ft. higher on its exterior than interior, and on the south is a secondary bank forming a fosse which is continued round the keep. The bank round the eastern court is much slighter. In the centre of this court there is a depression, evidently the remains of a pond, and at the far end of it are two others. There is a spring at the western end near the base of the keep. At the bottom of the field adjoining this court on the south are the remains of banks called 'The Fishers,' evidently the remains of fishponds. The natural level of the eastern court is lower than that of the western. The entrance would appear to have been from the north near the Keep.

Middleton Stoney.— In the park, near the church, is a pear-shaped mound, which has been cut into at its narrow portion situated north-north-east. It is 13 ft. 3 ins. high at its highest point, and on the east is a court enclosed by a bank and ditch, the latter being outside, and 5 ft. 9 ins. below the top of the former. There are no traces of a continuation of this on the west, where the church enclosure and a modern road have been made.

Mixbury: Beaumont Castle.—This work consists of two courts, the northern of which is regular in shape and practically square, while the south bank of the southern conforms to the course of the adjacent road which runs between it and the church. It is significant to note that this road is part of a footpath which runs from Brackley through Evenley Park, where Roman remains have been found, to the Roman Way which runs from Alchester to the north-east and which it joins at Finmere, two miles east of Mixbury. It was probably therefore in existence before the construction of the castle. The western rampart of the northern court ends on the north in a circular mound, the flat top of which is 30 ft. across. This rises 9 ft. from the bottom of the fosse, but opposite it in the angle of the latter there is a further depression 5 ft. deep, which, however, is probably not ancient. The mound is 5 ft. above the interior of the court. The northern bank does not quite connect with it, giving an appearance of an entrance from the fosse. The main entrance from the outside is at the north-west corner of the southern court, and then into the northern court in the middle of its southern side, a raised way leading it over the fosse which separates the two courts. A short distance from the entrance the ground level of the northern court rises about a foot. The fosse which surrounds both courts is of an average depth of 9 ft. below the top of their banks and 5 ft. below the outside ground. The surface of the ground of the northern court is irregular. There is a well a short distance to the west of the fosse.

Oxford: The Castle.—This was originally a mount and bailey fortress, but all that now remains is the mount, about 65 ft. high above the lowest point of the surrounding ground, which is a conspicuous object by the side of New Road adjacent to the county hall. On this stood the keep of the later stone castle and a stone well-chamber is enclosed below the top of the mound. The site has been so obliterated by the building of the county hall and gaol and streets that it is almost impossible to trace the course of the ditch enclosing the bailey with certainty, but its lines may be fairly conjectured. (fn. 12) On the south-west runs one of the many branches of the river which serves here as part of the Oxford Canal, and this may have formed the defence of the castle at this point, the broad ditch commencing on the west of the mound and, encircling the court, joining the stream again on the south. The ditch is now entirely filled up. From the conjectural plan thus laid down it would appear that the original fortress consisted of a mound with a single court on the south-east. The ditch was of unusual width, but was possibly widened to increase the defence of the castle, which lies very low, at a later period.

Swerford.—Here is a most interesting little specimen of this type of fortress. Indeed, one might almost be tempted to look upon it as a combination of the classes 'D' and 'E,' or to speak perhaps more accurately, of the development of the latter from the former. It consists of one circular court, the mean diameter of which is 150 ft., surrounded by a ditch, the level of the court being 16 ft. above the bottom of the latter, which is 6 ft. below the level of the surrounding ground. The court is, therefore, really a mound itself, the summit of which is surrounded by a bank 4 ft. high. On the north, however, where was apparently the entrance, is a smaller mound, the diameter of which is 50 ft. north and south, and which is tapered in shape to join the bank of the court on the east. It rises 8 ft. 6 ins. above the surface of the court. If we look upon the main court as an original mound this looks like a subsidiary one guarding the entrance. On this side the natural ground slopes to a brook about 100 yards distant. On the east are two detached raised platforms, beyond which the ground drops rapidly to the road. The larger of these carries a small mound, 12 ft. in diameter, the top of which is 6 ft. above the bottom of the fosse separating it from the mound at the north of the main court. Its height on the opposite side is 5 ft. 6 ins. above the platform, which owing to the natural fall of the ground at this point is itself 5 ft. high. The church stands on the edge of the ditch on the south, and the latter is rather shallower here, but this may be due to a subsequent filling in. The gardens of the houses on the west and the churchyard on the south have cut into the outer bank of the fosse on these sides.

Simple moated enclosures

(Class F)

Ardley.—In Ardley Wood, hidden amid the thick growth, is an ovalshaped enclosure surrounded by a ditch, the average depth of which is 3 ft., enclosing the site of Ardley Castle, now demolished.

Aston Rowant.—At Chalford hamlet in this parish is a rectangular moat enclosing a building, known as Cop Court.

Barford St. Michael. —A moat on three sides of the Manor Farm. It would appear to have formerly extended on the south side from a made-up bank and a feeder from the River Swere. (See plan.)

Brightwell.—A strip of water abutting the north end of Caldwell Farm buildings may represent the remains of a moat.

Britwell Salome.—A straggling piece of water on the north of the site of the church and castle (supposed) now both destroyed, may have enclosed them in its complete form.

Broughton.—Complete moat surrounding the Castle.

Bucknell.—Remains of moat at Manor House.

Chalgrove.—Remains of an irregular ditch enclosing the ground in front of Manor Farm.

Chinnor.—A broad strip of water on the north side of Henton Manor Farm represents the remains of a moat.

Chislehampton.—Remains of irregular moat surrounding site of Camoise Court, formerly a nunnery.

Clanfield.—Moat surrounding Friar's Court, on the site of a monastery.

Clattercote.—Small remains of moat at farmhouse on the site of St. Leonard's Priory.

Cropredy.—Remains of moat at Prescote Manor.

Curbridge.—There is an interesting series of three moats at Caswell House in this parish. The first is at the house and consists of four sides, having the appearance in the plan of a parallelogram opened out at the north-east corner, in which open space are the present buildings which the moat does not surround. At a distance due south of 170 yds. is a second moat consisting of three irregular sides, also with the open side to the north-east, and on this side until recently stood some buildings. Close to the south of this is a smaller and very regular four-sided moat, called Black Moat, with the entrance over the northern side, and now also devoid of buildings in the interior. From this a dry ditch leads on the south to the Norton Ditch, a small stream which feeds the first and third of the series. There is a well on the open side of the second. (See plan.)

Cuxham.—Slight remains of a moat at Manor Farm.

Filkins.—Three sides of rectangular moat at Moat Farm.

Fringford.—A very slight trace of a moat here at the Manor House, where the name 'Moat Gardens' also perpetuates it. There are likewise remains of three sides of a moat at Cotmore Farm in this parish.

Godington.—A rectangular moat with a short continuation on the north side towards the east and completely enclosing the house and out-houses with entrance from the south. (See plan.)

Henley.—A straight piece of water on the north of the site of Phillis Court on the banks of the Thames may be part of a moat.

Holton.—A rectangular moat on the site of Holton House, now destroyed.

Remains of a moat on the east side of Church Farm in this parish, near the road at the north-east corner of the park.

Islip.—Remains of a rectangular moat on the supposed site of Ethelred's Palace.

Kirtlington.—In the centre of the village, in an enclosure between the school and the park wall, is a moat somewhat in the shape of a triangle, 26 ft. across at the narrowest part at the north angle and about four times that width on the east, enclosing an area of about 150 ft. diameter. It is marked on the Ordnance Survey Map as the supposed site of John of Gaunt's castle.

Lewknor.—A rectangular moat completely surrounding the buildings known as Moor Court, with entrance from the east. (See plan.)

Newton Purcell.—Remains of a moat on the east of the village on the site of the house of the Purcells, the buildings of which have been long destroyed.

Northmoor.—A complete rectangular moat with house and buildings inside at Manor Farm.

A ditch fed from the River Windrush encompasses the whole block of buildings and grounds known as Gaunt House in this parish.

Oddington.—A large irregular ditch west of Oddington Grange, enclosing the site of a monastery.

Pyrton.—On the north of Golder Manor is a small rectangular moat enclosing a space now destitute of buildings.

An irregular piece of water resembling a pool, but called 'The Moat,' is on the east side of Pyrton Manor House.

Shirburn.—The castle is still perfect and stands in the centre of a moat, rising straight from the water. It is reached by three bridges over the moat, which on the north side is much broader, and is carried 50 yds. further westward into the grounds.

Stanton Harcourt.—The Parsonage Farm is surrounded by a moat on the three sides farthest from the village, at the north end of which it stands. On those sides also the garden is enclosed by a secondary moat beyond the one round the house, while a ditch would appear to have formed a third enclosure between these two and the stream running on the west of the house.

Stanton St. John.—A planted enclosure called 'The Moat,' in a field south-west of Studley Wood, would appear to cover the remains of a rectangular work.

Stratton Audley.—A plain rectangular moat enclosing the site of the castle now destroyed, with entrance from the north-west, remains in a field called 'Court Close,' east of the church. The work lies on low ground in the bend of a stream and about half a mile to the east of the Roman Way from Alchester, which is supposed to explain the first part of the name of this village. This situation is, perhaps, worth noting in view of the suggestion that these moated enclosures may mark the sites of more ancient works, and also in this respect is the fact that its plan corresponds as regards its relation to the compass points with those of the rectangular enclosures in this county supposed to be ancient (see under class 'C'), its sides running south-west to north-east and north-west to south-east. (See plan.)

Thame.—Remains of moat at prebendal chapel.

Watlington.—A ditch on the south side of the site of the castle is supposed to be the remains of its moat.

Weston-on-the-Green.—At the Manor House are two sides of a rectangular moat of which the third side can be traced, and which would enclose the space in front of the house, the house standing on the open side.

Wychwood.—At High Lodge there is a small moat enclosing the house and interesting from its situation, which, before the disafforesting, was in the centre of Wychwood Forest, and on the highest point of ground in the forest. It is a con spicuous object now as it stands on the top of a high hill which rises from all quarters, and which on the south ascends 200 ft. in three-quarters of a mile. The altitude of the moat is 631 ft. above sea level. (See plan.)

Yelford.—Rectangular moat south-east of Manor Farm.

On the sites of Langley Palace, Beckley Palace, and Somerton Manor House, in those parishes, are mounds which cover the remains of mediaeval houses now destroyed. They do not appear to have been moated. In the field in front of the rectory at Somerton are the banks of the fish-ponds. Neither of these remains fall within either of the classes of the works dealt with in this chapter, but it is, perhaps, necessary to notice them, as their grass-grown mounds give them the appearance of being earthworks.

Moated enclosures with stronger defensive works

(Class G)

Bampton: The Castle.—A raised court, 500 ft. by 200 ft., bounded on the east by the stream, to the north of the farm-house called Ham Court, which is the gateway of the castle, now demolished, is all that represents this work. On the west the ditch, which apparently enclosed a larger area, remains, extending southwards beyond the court to the gate-house, where it broadens to almost twice its former width.

Barford St. John. —A moat about 7 ft. deep and 56 ft. wide runs on the north and west of an area which is roughly the shape of an irregular pentagon, and which is divided into two courts by a raised platform about 2 ft. high in the northwest corner. This probably marked the site of the buildings, the remainder of the area being an open court, round the north and east edges of which is a slight bank. The moat on the south of the area is now of very slight proportions, but the ground is very low, and the standing water to the east is no doubt the remains of the defences which ran along this side, and were continued between the banks which remain to the west. These banks are 5 ft. high from the bottom of the ditch they form, which is 38 ft. across. On the north of these the ground rises, and on the slope is a small horse-shoe shaped work, evidently the site of an out-building. The site encloses no buildings at the present time. (See plan.)

Beckley.—At Lower Park Farm, in this parish, is a double-moated enclosure of a regular rectangular form, in which stands the farm-house. Of the outer moat the north-west, south-west, and north-east sides remain, and enclose an area 200 ft. square. The fourth side has probably been destroyed by the farm buildings. In the centre of this are the remains of two sides of a second ditch, the sides of which are parallel to those of the outer. The two which remain are the north-east and south-east, and they appear to be the remains of an inner enclosure 100 ft. square. In view of the suggested ancient origin of some of these moated enclosures, it is, perhaps, worth while to draw attention to the fact that the compass bearings of the sides of this work correspond to those of the ancient rectangular earthworks of this county, and to those of Stratton Audley Castle (q. v. under class 'F'), while its position in regard to the Roman Way is the same relatively as Stratton Audley. A Roman villa was situated on the south-west between this site and the Way. (See plan.)

Cogges: Site of Castle.—On the banks of the River Windrush, in this parish, is a double enclosure forming two courts, one to the south of the other. The southern one is entirely surrounded by an irregular four-sided moat, and the northern, which is of about equal area, has the river for its western side and a ditch on the eastern. On the northern side is a deep ditch with a rampart on the inside springing from the bank of the river and curving outwards. The entrance was apparently at the northeast corner of this court. The rampart is 2 ft. 6 ins. above the surface of the court and 8 ft. above the bottom of the ditch at the western end near the river, and gradually increases in height until at the eastern end it is 6 ft. above the court and 12 ft. above the ditch. The bottom of the ditch is 5 ft. below the level of the outside ground. The moat throughout is of an average width of 17 ft. (See plan).

Radcot.—At Radcot Bridge, on the Oxfordshire side of the Thames, is a moated meadow, the enclosure being 400 ft. by 500 ft. and resting on the bank of one of the streams of which the river consists at this place. In form it is more like those of class 'F,' but it does not appear to have been constructed to surround or defend a homestead or buildings of any kind, and as it is known as 'The Garrison,' it is probably of a military character, although we are hardly justified in looking upon it as ancient. It was more probably constructed at the time of one of the encounters of which this spot was the scene in the fourteenth and seventeenth centuries.

Moats and ramparts surrounding village sites

(Class H)

Nether Worton.—There are remains of a moat surrounding Nether Worton House, and also of a continuation of it to the north beyond the church, which is situated about 70 yds. to the north-east. Its course from the house round the church can be easily traced, but its further line has disappeared. It is most probable, however, that it included the Manor Farm, at the garden of which it now ends. The three buildings thus named as falling within the compass of the moat constitute the whole of the village. A meadow to the east is called Mill Close, and was probably the site of the mill which has been destroyed, and here there is a broad ditch which, with the stream, would form a water enclosure within which the mill very likely stood, and, if so, all the holdings of the village would be protected by water.

Shelswell.—Two small fragments of a moat remain near the Home Farm on the site of the Manor House, which from their direction may also have encompassed the church, now destroyed, in which case the moat would have included the whole holding.

Witney.—The Ordnance Survey Map marks the site of a 'Saxon Rampart' along the edge of the hill which falls abruptly to the valley on the west of the town. Such a work would appear from the map to have encircled the town on the south and west sides. There would appear, however, to be now no remains of it, unless the broad ditch of about 5 ft. deep on the south of the churchyard and vicarage grounds, and which is in the line of the course laid down on the map, represents it. This may, however, be a boundary ditch of a later time. At the foot of the hill, and running round this side of the town in a curve concentric with that marked for the rampart, is 'Emma's Dyke,' an artificial watercourse 12 ft. wide, but it would appear to have been constructed for drainage rather than for defence.

Unclassified earthworks

(Class X)

Checkendon: Castle Grove, Wyfold Wood.—In Castle Grove, south of Wyfold Court, there is a large earthwork, described on the Ordnance Survey Map as a moat, and consisting of a ditch 9 ft. wide at bottom enclosed by two banks, the inner rising 2 ft. 3 in. from the level of the area surrounded by the ditch and 6 ft. from the bottom of the ditch, from which the outer bank rises 4 ft. and is about one foot higher than the ground outside. The ditch takes the shape of an irregular octagon, but the north side is open. Here in the centre of what would have formed the eighth side is a small pond, and the surrounding ground is marshy, and may have formed a natural defence. The entrance appears to have been on the south-west. The surface of the interior is level. The natural ground slopes away to the west, but is level on all other sides. (fn. 13)

Stratton Audley.—On sloping ground, with higher ground above it, on the west side of Oldfield's Copse and by the side of the road one and a half miles north-east of the village of Stratton Audley are the remains of a small circular earthwork. It consists of a bank enclosing a flat depression in the ground of a diameter of 93 ft. On the western side where the ground is highest this bank is only a few inches high on the outside, but it falls, in a slope of 48 ft., 6 ft. 9 ins. to the bottom of the interior of the work. The bank on the other sides is only about 1 ft. high. The entrance was apparently on the north. It is commonly called a Roman camp, and the Roman Way from Alchester passes at a distance of three-quarters of a mile to the west, but it presents none of the features of a Roman work, and its situation and small size render it of very little value for defensive purposes. An old inhabitant of Stratton Audley told me the banks had been much lowered in his time, and that rusty blades had been found there. Locally it is known as 'Stuttle's Bank.' It has none of the characteristics of an ancient earthwork, and its present appearance suggests a dried-up pond more than anything, but tradition associates it with soldiers, and the finding of the blades, which were probably weapons of some kind, supports the tradition, and it may have been thrown up for temporary purposes by a small body of troops.

Tadmarton.—In Tadmarton parish and threequarters of a mile due east of Madmarston Camp is a line of bank running west-south-west and east-north-east for a length of 450 ft., and being 5 ft. in height. It is grass-grown and planted with bushes. The ground to the south is ploughed, but not that on the north. At each end the ground suggests that the bank turned northwards, but there is now no trace of its continuation. A hedge runs parallel to it on the north at a distance of 150 ft., separating the field in which it stands from another which is under the plough, and in which there is now no trace of any work. There is no trace of a ditch on either side of the bank.

Defensive earthworks now destroyed

Eynsham.—East of Eynsham Park in an arable field, at a spot 200 yds. due west from the junction of the road from Eynsham village with Cuckoo Lane at Tanner's Hill Clump, was a small earthwork which appeared on the Ordnance Survey Maps before the last revision, from which it would appear to have been oval in shape with the longer axis running north and south 250 ft. in length and the shorter 225 ft., the western side being somewhat irregular.

Hook Norton.—Half a mile south-west of the circular camp on Tadmarton Heath, and a quarter of a mile west of the cross-roads on Wigginton Heath, and just inside the hedge of the second field from the road near an old quarry, was a small earthwork now completely destroyed. It was in shape an irregular pentagon, the dimensions of which are given in Beesley's History of Banbury as follows: West side, 52 yds.; south, 69 yds.; east, 38 yds.; north-east, 39 yds.; north-west, 63 yds.; total circumference, 261 yds. It was described as then (1841) reduced nearly to the level of the soil. Excepting on the south he found traces of an outer vallum. Sufficient remained for this work to be plotted by the Ordnance Survey in 1881, but in the revised map of 1900 the site of the work is marked entirely by dotted lines.

Kiddington: Hill Wood.—Warton in his History of Kiddington (1782) says that in this wood is a square Roman entrenchment with the ridge and fosses in extraordinary preservation, and as yet undetected by any topographer. Hill Wood is a very thick plantation, the dense undergrowth of which makes it extremely difficult to properly survey the ground. No earthwork is marked here by the Ordnance Survey either in their 1834 or 1881 maps, and I have searched the wood and found nothing in the way of banks and ditches but the boundary banks which mark the delimitations of the different copses in all these woods. The wood covers the slope of a steep hill which falls in a sharp descent towards the north-east, parallel to the road to Oxford south of Over Kiddington, and is shown on the map of the course of Grim's Ditch 'A,' which runs through an adjoining coppice to the south-west, and if such a work exists it will probably be found in the small portion of the wood on the top of the hill and to be of the character of the miniature camps described under Cornbury Park and Spelsbury in class 'C,' also on the course of the ditch.

Spelsbury.—The 1834 edition of the Ordnance Survey Map shows a rather large rectangular camp on the west of Ditchley Park on the Grim's Ditch, which appears to form its northern side, at the spot where stands the Model Farm. There is now no trace of this, and it is not shown on the later maps.

Spelsbury: 'Castle Ditches.'—In 1842 Alfred Beesley, the author of the History of Banbury, was informed that there was 'until lately' a camp at Spelsbury. (fn. 14) Lewis (fn. 15) says:—'On an eminence near the village is an extensive triangular entrenchment called Castle Ditches, enclosing a space of about 24 acres.' It would appear to have been recently destroyed when Beesley received his information.

Stonesfield: Callow Hill.—On the main road from Charlbury to Woodstock and 3½ miles from the former place is marked on the 1834 edition of the Ordnance Survey Map 'Callow Hill,' in the distinctive lettering used for ancient sites, which does not appear on the later maps. Warton in his History of Kiddington says that there was a small Roman encampment here, and General Pitt-Rivers records that there was a rectangular enclosure thickly strewn with Roman tiles and pottery. (fn. 16) I can find no more exact description of the work. The site would appear to be at the present north-east corner of King's Wood, and there is now no trace of the enclosure unless it be represented by the modern hedge which encloses a roughly rectangular space of about 500 ft. by 150 ft. with a quarry at the eastern end. It is, however, doubtful whether the work was of a defensive nature, and it is more probably the site of a Roman villa, of which there are several other remains in the immediate neighbourhood.

Dykes, Banks, etc.

There are in Oxfordshire remains of two extensive lines of entrenchments known as Grim's Ditch. One is in the middle of the county and the other in the extreme south. The former I call 'A' and the latter 'B.'

Grim's Ditch 'A'

This is a bank and fosse which runs in a semi-circular course, commencing in the northern portion of Blenheim Park and passing through some portion of the parishes (though not in any case touching the villages) of Glympton, Kiddington, Enstone, and Spelsbury, and ending, so far as its present remains have been found, on the east bank of the River Evenlode, south of Charlbury town. Dr. Plot (fn. 17) in 1676 and Dr. Warton (fn. 18) in 1782 traced it, but it is difficult at points to follow the course as laid down by them. The work exists at present in fragments, and first appears in Blenheim Park in the form of a slight entrenchment at the place south of the Ditchley Gate, where Akeman Street crosses the park. There is now no trace of it as it leaves the park, but 650 yds. due north of the gate it reappears in the further hedge of the field in which Woodley's Farm buildings are situated. From here it runs in a northerly direction down the hill, which falls 68 ft. in 500 yds., to Slape Bottom. It consists of a ditch 38 ft. wide, with a bank along the western edge the top of which is 6 ft. 4 ins. from the bottom of the ditch and 2 ft. 9 ins. from the ground level. The ditch is 4 ft. below the ground level on the eastern side (see section A—B on map). At Slape Bottom it enters Hark Wood and ascends the opposite hill amid the woods which clothe its side. Here it is not so pronounced. About 80 yds. from the bottom of the hill a barrow rises in its course, after which to the northern end of Hark Wood it exists in the form of a bank 7 ft. 6 ins. from the bottom of the ditch, which is to the east, and is 43 ft. wide and 3 ft. 6 ins. below the ground, which here slopes to the east. All traces of it are lost at the end of Hark Wood, but it begins again suddenly with a dead end due apparently to no natural feature at a point in Berring's Wood 570 yds. north-west of its end in Hark Wood and 130 yds. south-east of Tomlin's Gate. Here as before the ditch would appear to still face eastward, and is of the same width as before, being 6 ft. 6 ins. below the ground on the west and 3 ft. 6 ins. on the east. From the top of the western bank the ground slopes with a natural declivity to the rear; in fact the ditch runs along the eastern side of a defile, shown on the Ordnance Survey Map by the 400-ft. contour line, for about 540 yds., when it turns westward, and after crossing an open field, where its dimensions have been very much reduced, enters Out Wood, in the middle of which it ceases. There is no further trace of it in the direction in which it was pointing, but on the summit of the hill, 600 yds. north of Out Wood and at an altitude of 449 ft., it re-appears in a small section in the field to the west of Assarts Cottage. It runs due east and west, and has been practically demolished by the plough, but appears as a small bank facing north. It is to be seen again at the western end of the same field, and can be traced, still in a very much reduced condition, across the Kiddington Drive of Ditchley Park. In the park there are no traces of it until a quarter of a mile to the west of the house near the Rosary, where it appears again with the ditch 40 ft. wide. It has here the local name of 'Love Walk,' and runs east and west on the north of the road from Ditchley House to Model Farm, at one of the western entrances to the park. At the farm it makes a curious sweep to the north as if to skirt the buildings, and comes to an end on the road, which is apparently an old trackway. There is no raised bank remaining along its course at this place, but the southern bank of the ditch is 6 ft. high from the bottom, while the northern is only 3 ft. 6 ins. The ground here slopes naturally to the north. At this point the ditch would enter the cultivated and open ground between Ditchley and Charlbury, and there is no trace of it until it re-appears in a field called Baywell to the south of the latter place, where it runs in the shape of a bank 4 ft. high with the remains of the ditch on the north along the northern side of a natural gulley, and comes to an end above the railway and the River Evenlode, pointing straight for one of the rectangular earthworks in Cornbury Park on the other side of the narrow valley. This is the last appearance of it so far as it can now be traced, but Dr. Plot asserts that he was told that it could be found in the woods beyond Cornbury Park pointing for Ramsden. In this case it would probably again join the Akeman Street, which runs through that parish, and from which it started in Blenheim Park five miles eastward. The Roman road, if this were so, would form a base upon which Grim's Ditch, as sketched above, would form a semi-circular arc to the north, its centre being at Stonesfield, near which village important remains of Roman villas have been found.

The remains of two other pieces of entrenchment must be mentioned as apparently forming part of the scheme of Grim's Ditch. The first is a trench which commences at Starveall Farm, and, crossing the Charlbury to Woodstock Road, runs due north down the hill to Pool Bottom, parallel to but about half a mile to the west or in the rear of the section of Grim's Ditch between Woodley's Farm and Slape Bottom. This ditch, which is on ploughed land and is being rapidly reduced, is now 70 ft. from bank to bank, the westward being 3 ft. high and the eastern half that height.

The second of the remains is of a short line of entrenchment which formerly ran east and west at 100 yds. north of Shilcott Wood above Ditchley New Park and parallel with that section of Grim's Ditch from Ditchley Park to the Model Farm. It has disappeared in the fields, but can still be traced where it crosses the old trackway which runs down the west side of the park and which at this spot is a grass-grown lane. This entrenchment would be in front of the Grim's Ditch assuming that the latter faced outwards, as the different heights of its bank would appear to justify us in looking upon it as doing.

In 1868 General Pitt-Rivers examined the ditch and the ground encompassed by it, and came to the conclusion which he stated as follows:— (fn. 19)

I have only to add, from personal inspection of it, that it is not merely a boundary, but without doubt a fortification, for its commanding position, its adaptation to the features of the ground, and the situation of its ditch, are points which, viewed tactically, are sufficient to determine it to be a work of defence. Throughout its whole line it so much resembles other dykes which I have examined in Yorkshire and elsewhere . . . that if I were to be guided by its trace alone I should be inclined to class it with those dykes and to attribute it to the same origin, but other considerations are favourable to its being a Roman earthwork. These considerations are: firstly, that it covers a portion of the Akeman Street, which runs across the country in a north-easterly direction, passing along the rear of this work in such a manner as to be defended by it from a northerly attack; and secondly between Akeman Street and the Dyke, and within the area defended by the Dyke, there are traces of several Roman villas and other Roman remains. . . . These circumstances favour the supposition that the Dyke may have been thrown up by the Romans to defend a Roman settlement established in this place in connexion with the great road and to secure the communication of the inhabitants with the road.

Against this assumption of a Roman origin must be set the evidence of Mr. F. Haverfield, who in a communication on his examination of the Akeman Street in Blenheim Park in 1898, (fn. 20) points out that so far as could be judged the road ran over the ditch and cut through the entrenchment (which is taken to be the commencement of Grim's Ditch), and if this be so the entrenchment at that point, at all events, would appear to be pre-Roman.

Grim's Ditch 'B'

The second dyke bearing this name in this county appears to have run from the River Thames opposite Wallingford in a south-easterly direction across the Chilterns to the same river at Henley-on-Thames, the distance between the two points being ten miles. The name Grim's Ditch applied to this work appears in a charter of or before the reign of Richard I. Commencing on the west the ditch appears to have run from the river bank three-quarters of a mile below Wallingford Bridge, along the north side of the grounds of Mongewell House to the road from Crowmarsh Gifford, where it reaches higher ground. It then runs east-by-south in a straight line for 3 miles to Nuffield, and this section is the most perfect of its whole course. From the lodge of Mongewell House to the edge of Foxberry Wood, where it crosses the Icknield Way, about a mile and a quarter, it exists in the shape of a grass-grown bank 5 ft. 9 ins. in height, with a flat top 4 ft. wide which appears to be used as a footpath. It runs here across high ground which slopes gradually north and south, though more so to the latter, and as the land is arable it forms a conspicuous object seen for some distance when the ground is clear of crops. In Foxberry Wood and Oaken Copse the bank had been reduced, and outside the latter to the east it disappears altogether for about 200 yds. in the course of a modern hedge. Then it re-appears in the shape of a ditch and enters the belt of trees leading into the Mongewell Woods, and here the bank appears again on the north side of the ditch. As it descends into Morrell's Bottom and the wooded enclosure north of Mongewell Woods the bank rises 6 ft. from the ground level on the north and falls 8 ft. 6 ins. into the ditch, which is 7 ft. below the ground level on the south and is 45 ft. wide from bank to bank. Inside the enclosure the bank disappears and the work continues in the shape of a ditch only: at first, near Woodlands Farm, with the sides of equal height, but nearer Nuffield the northern side is 9 ft. 6 ins. above the bottom of the ditch, and the southern 4 ft. 6 ins. It is here pointing straight for Timber's Barn, but about 6 yds. inside the field in which the barn stands it comes to an end, to re-appear again in the form of a ditch on the northern side of a rectangular wooded enclosure called Heycroft Shaw, at the south corner of Nuffield Common and half a mile due east of its end near Timber's Barn. It did not, however, apparently go straight across this interval, for a bank about 270 yds. to the south would suggest that the Ditch made a dip to this point and then turned north-east to its present trace at the corner of Nuffield Common. Continuing from the latter point a bank running south-east carries it to Hayden Farm. The remainder of its course to Henley exists only in fragments, and from these it appears to have been of a rather tortuous character. It is a significant fact that down to Lambridge Wood, where the last remaining section of it is to be found, the fragments of it are all on the line of the southern parish boundary of Nettlebed and the western and southern boundaries of Bix. It is not an unreasonable suggestion that these parish boundaries were laid along the Ditch, and that in those parts where it has disappeared the boundaries mark its course. It is upon this that the conjectured line of its course has been laid down on the map herewith. Often these dykes became used as footways and obtained in places the local names of lanes. In the Grim's Ditch 'A' a footpath runs down it from near Woodley's Farm to Slape Bottom, and near Ditchley Park a section of it retains the name of 'Love Walk.' A continuation of a fragment of the Ditch under present consideration which appears at the north of Swan Wood is a lane called 'Deadman's Lane,' while a portion of the southern boundary of Bix, in one of the gaps of the remaining course of the ditch, but leading straight to the one in the woods above Greys Court, runs along a lane called 'Rocky Lane.' In these lanes we may possibly see further evidence of the course of the Ditch, and if so they support the suggestion that it was used in the demarkation of the parish boundaries.

Reverting to the remaining traces after Hayden Farm in the parish of Nuffield, a footpath along a hedge planted with trees and sweeping in a curve first south and then south-east would appear to be its course. The boundary of the Nettlebed parish joins this and runs along it, and upon it at the northern edge of Swan Wood there is a trace of the Grim's Ditch in a bank from which runs the lane called 'Deadman's Lane.' The present boundary leaves this to make a straight turn to the north-east for 120 yds., and then at right angles again back to the lane, and the latter would appear to have been the more probable continuation of the Ditch, which re-appears at the point where the boundary rejoins the lane and runs due south, skirting the western edge of Highmoor Common Wood, for 700 yds. and then turns, still carrying the parish boundary, due east for 300 yds., where it ceases. In Highmoor Common Wood, at the point where Grim's Ditch re-appears on the north, there is a trench running to the east through the wood, known as Highmoor Trench, which can hardly, however, have been part of the course of the Ditch. After leaving the remains of the Ditch at the south of Highmoor Common Wood the parish boundary continues eastward to a plantation called Broom Pightles and then southwards to 'Rocky Lane' and eastward along it into the woods north of Greys, in which slight remains of the ditch have been traced on its course, and then up to the north-west corner of Lambridge Wood. There the Ditch re-appears, but leaves the boundary and in the form of a ditch makes a sweep to the south-east for about half a mile, and then comes to an end, pointing straight for Henley town. This is the last trace of the Ditch.

Ash Bank, Wattle Bank, or Avesditch

This was an entrenchment running north and south on high ground to the east of the River Cherwell and about two miles from it, for a distance of about six miles, from just below Souldern on the north to near Kirtlington on the south. Very few traces of it now remain, and these do not suggest that its dimensions equalled those of the Grim's Ditches. Its line, commencing at the north, appears to have been along the road from Inkerman Farm, south of Souldern village, to Fritwell, through that village and along the course of the lane called Raghouse Lane, by Kennel Copse and Ballard's Copse (the ancient name of which was Chilgrove), to the Heath in the parish of Middleton Stoney. Of its course so far, which was almost due south in direction, only a very few slight traces now remain. At the Heath it turned south-south-west and can still be distinctly traced along the whole length of the large meadow to the west of Middleton Park. Here it is now a ditch 27 ft. wide, with the eastern bank 4 ft. 9 ins. from the bottom and the western bank 3 ft. 6 ins. The length of this traceable section is a mile and a quarter, and on the western side runs a cart-road which at the south-west corner of the field takes the form of one of the old enclosed trackways which no doubt it was originally throughout. The ditch, however, disappears at this point, but threequarters of a mile further on and a mile north of Kirtlington the trackway joins the Portway, and here the ditch re-appears and crossing the road maintains the same compass bearing that it had before, and is to be traced for 630 yds. in the fields, when it ends abruptly in the middle of the second field, pointing straight for the spot where the Akeman Street crossed the Cherwell. These fields are ploughed and the ditch is being rapidly filled in by the operation, but the bank facing north-west is still distinct and has a height of 18 ins. above the ground at its foot. The Portway, generally called a Roman road, which the ditch crosses above Kirtlington, runs between it and the Cherwell, and appears from its direction as though it may also have joined its northern end near Souldern.

Local Dykes, Banks, etc.

In addition to the three entrenchments described above, which cover several miles of country, there are four smaller ones, which are distinctly local in character.

North Leigh.—In this parish in the fields half a mile south-west of the village and commencing in the second field from the junction of the road from the village with that from Witney on the opposite side of the road to Eynsham Park, is a line of entrenchment running east and west for 650 yds. down the side of a hill sloping westwards. It consists of a slight bank with a ditch on the south 5 ft. 3 ins. below the top of the bank and 18 ins. below the level of the ground beyond.

Bicester: Gravenhill.—At the eastern extremity of the wooded hill which rises so conspicuously from the level country to the south of Bicester, and which is half a mile due east of Alchester, there is a line of entrenchment which follows the top of the hill for 200 yds., along its northern face and then turns along its south-eastern edge for 380 yds., making an acute angle at its eastern point. Along the northern face it would appear as though the natural fall of the ground had merely been emphasised by scarping. On the south-eastern face the work consists of a ditch 2 ft. 3 ins. deep from the higher bank above and rising 15 ins. in front. The entrenchment would appear to be a local and temporary one to meet an attack from the east, from which direction Akeman Street approaches it, rather than a general fortification of the hill top.

Swyncombe: The Downs.—Along the top of the northern face of the bold spur of the Chilterns called Swyncombe Downs runs an entrenchment commanding the Icknield Way, which approaches the base of the hill from the north-east, and the large expanse of plain from which the hill rises steeply some 300 ft. It consists of a trench 4 ft. deep and 35 ft. wide. It commences near the highest point of the hill on the east, and running due west along its northern edge turns north-west and descends the slope to within 130 yds. of where the Icknield Way crosses its base. Its total length is one mile, and from the point where it turns north-west to descend the hill towards the road there appear to have been two slighter lines running to the rear of it to the west and south-east, as though to protect this narrow part of the hill nearest the Way from being taken in the rear.

Witney.—In a field on the south side of the road from Witney to Burford between the first milestone from Witney and the road to Crawley, is a low bank just within the hedge of the field and running parallel to it for nearly its entire length and then turning to the south-west. It has the appearance of being a boundary bank.

Tumuli, Barrows, etc.

Long barrows

Lyneham.—About 350 yds. south-east of Lyneham Camp on the south of the hill and commanding extensive views on all sides but the north. The major axis runs north-north-east to south-south-west, and appears to have been 184 ft. long and 44 ft. wide at the broadest part, but the reduced state of the northern end, about a third of the whole, makes it difficult to give an exact estimate of its original dimensions. The wall of the field runs across the top of the barrow about 80 ft. from its northern end, and beneath the wall there is a shallow depression which would appear to suggest that the barrow had previously been opened at this spot. The portion to the north of this wall is being rapidly demolished by the plough. The portion south of the wall is grass-grown and well preserved, the highest point being 7 ft. from the ground level. On the top of the barrow, 40 ft. from the northern end, stands a monolith, the height of which above the reduced surface of the barrow is 5 ft. 9 ins. It is 5 ft. 10 ins. wide and 18 ins. thick.

The barrow was opened under the direction of Mr. Edward Conder in 1894, and he contributed a paper giving an account of the excavations and finds to the Society of Antiquaries, (fn. 21) accompanied by a plan and sections of the barrow. It was found to be constructed of the rubble stone of the district. Several trenches were sunk and the finds consisted of a long cist on the level of outside ground, under the floor of which were a certain amount of animal charcoal, a few fragments of bone, a tooth, some fragments of lightly baked dark-coloured pottery, one piece of which was marked with white lines; a second and smaller cist slightly above the ground level with a perforated floor stone; above this and about 9 ins. below the surface of the barrow fragments of a human skull and a portion of the humerus; a complete human skeleton lying north-east to south-west, apparently a later interment, a javelin 9¼ ins. long characteristic of the Saxon period and a knife 8¼ ins. long close to the right of the skeleton; a skeleton lying full-length nearly north and south with a small knife 5⁷/8 ins. long; calcined stones; fragmentary remains of a third skeleton close to the surface lying on the left side in an L-shaped trench, head to north and feet to east; a large monolith 4 ft. 6 ins. by 3 ft. and 11 ins. thick in a horizontal position at the north-west end and 2 ft. below the surface, lying east and west, at which latter end it touched a line of carefully laid roughly quarried stones; a skull on the ground level resting on two flints with a third flint in close proximity; a quantity of human bones and the fragments of at least four skulls on ground level and all heaped together; remains of small knife and umbo of Saxon shield; four pieces of iron which pieced together measured 20 ins. long, 2 ft. below surface; skull and remains of the ulna of a child, horse's teeth and signs of fire slightly below ground level; in a dug-out hole 9 ins. below ground level some small flints, fragments of human bones, pieces of wood charcoal and traces of burning. The total height of the standing monolith proved on excavation to be 10 ft. 6 ins. The opinion of the meeting was that the primary burial had not been found, and it was suggested that it may have been at the spot where the wall crosses the barrow and where the depression of its surface suggests a previous opening.

Wychwood Forest.—In Slatepits Copse is a long chambered barrow, at a point 160 yds. south of the road leading through from west to east and 67 yds. on the east of the road from north to south. It is not marked on the Ordnance Survey Map. It lies on ground sloping to the south, its long axis running east-south-east to north-west-north, and is broader and higher at the eastern end, where the remains of a small stone chamber have been exposed. The length is 100 ft., and the width at the broadest end 56 ft., and at the narrow end 44 ft. The height at the eastern end is 5 ft. 6 ins. above the ground level on the north and 6 ft. on the south, where the ground is naturally lower owing to the fall of the hill. The chamber is a little to the north of the centre line, and 20 ft. from the eastern extremity. There are three stones exposed, two parallel with the length of the barrow, the exposed portion of the larger on the south being 5 ft. 6 ins. above the ground and 6 ft. wide; and the smaller, which is 5 ft. to the north of this, 2 ft. above the ground, and 1 ft. 6 ins. wide. These formed the parallel walls of the opening, and the third is across the rear of them and slanting forward, the exposed portion measuring 4 ft. in height and 5 ft. in width. At the western end the height of the barrow is 2 ft. 6 ins. above the ground on the north, and 5 ft. 6 ins. on the south. It has been disturbed. Akerman, (fn. 22) who is, so far as I find, the only writer who has noticed this, apparently overlooked, example of a chambered barrow in this county, says it was plundered a few years before he wrote by a keeper. One of the present keepers told me they had dug into it for rabbits and had found nothing.

Round barrows

Adwell.—On the summit of Adwell Cop, well preserved, 350 ft. circumference, 12 ft. high.

Asthall.—On the south of the road from Burford to Witney at the point where the road from Asthall joins it, planted and well preserved except that its shape has been somewhat spoilt by a wall which has been built round its slope to a height of 4 ft. 6 ins., 190 ft. circumference, 8 ft. 6 ins. high.

In a field on Leigh Hale Plain in Asthall parish, west of Stockley Copse, 215 ft. circumference, 5 ft. high.

Badgemore.—'The Mount' in Henley Park.

Black Bourton.—About half a mile south of the village on the west of the Faringdon road, now partially destroyed by the plough.

Chadlington.—Near Jubilee Plantation, in ploughed field, almost destroyed.

South-west of Old Downs Farm, almost destroyed.

Chinnor.—On the summit of Chinnor Hill, a northern spur of the Chilterns 800 ft. in altitude, and rising 300 ft. from the Icknield Way which runs at its northern base, a pair of twin barrows enclosed in one ditch, 200 ft. circumference and 6 ft. high respectively.

Another on hill summit in Chinnor Wood, 175 ft. circumference and 4 ft. high.

Churchill.—On the south side of Borsbery Lane, planted with trees and surrounded by a wall and in good preservation, 200 ft. circumference and 9 ft. high.

'The Mount,' on high ground on the south of the village, surrounded by a ditch 3 ft. below the ground level, in admirable preservation, grassgrown and planted, 260 ft. circumference, 12 ft. 6 ins. from the ditch in height.

Cornbury Park.—There are two barrows in Cornbury Park close to the small earthworks described in class 'C.' The northern one is 100 ft. circumference, 1 ft. 6 ins. high; the southern 200 ft. and 6 ft., with a very old oak growing from the top. The northern appears to have been opened.

Crawley.—There were five barrows in the fields bounded by Riding Lane on the west, Pay Lane on the east, and Akeman Street on the south. One near Chasewood Farm had disappeared when the Ordnance Survey map was revised in 1898. Two of the other four have since been destroyed, the remaining pair being close to the hedge on the south of Blindwell Wood. One on the south of the hedge is on ploughed land and is being rapidly reduced. The other on the north side of the hedge is planted, and although better preserved has evidently been mutilated. It is 115 ft. circumference and 5 ft. high.

Enstone.—In ploughed field on bridle road from Lidstone to Spelsbury Down Farm, almost destroyed.

Mound in Ditchley Park near Kiddington Drive in Enstone parish, surrounded by a ditch 4 ft. 6 ins. deep, 343 ft. circumference, 9 ft. high from bottom of ditch.

Glympton.—In Hark Wood and on the Grim's Ditch, 160 ft. circumference, 6 ft. high.

Kiddington.—In the field north of Wood Farm, grass-grown, planted and well preserved, 230 ft. circumference and 6 ft. high.

In King's Wood Brake, 165 ft. circumference, 5 ft. 6 ins. high.

Langley.—Two small barrows east of Hen's Grove. The southern of the pair has been opened by a trench to its base and is left in that condition. It is constructed of rubble stone, which appears to extend 1 ft. below the ground level at the centre, 175 ft. circumference, 4 ft. high. The other has been almost levelled by the plough.

Leafield.—On high ground to the north of the village and commanding an extensive view in all directions, grass-grown, planted and well preserved, 320 ft. circumference, 11 ft. 6 ins. high on west and 8 ft. on east. It has the appearance of having been opened.

Lew.—On summit of a hill on west of village, planted and well preserved, 350 ft. circumference, 9 ft. high.

Mixbury.—On Barrow Hill, which is only a slight rise in a field on the west of the village, and on which the barrow was constructed. Much reduced by the plough. Situated close to a hedge on the other side of which the Ordnance Survey marks the discovery of human remains.

Over Norton.—Mound on the summit of the hill in Over Norton Park, 161 ft. circumference, 3 ft. 6 ins. high.

Over Worton.—On the hill on which the church stands and just outside the north wall of the churchyard is a mound 198 ft. circumference and 9 ft. high, planted, and having the characteristics of the other round barrows of this district. A report in the parish says that it covers a heap of rubbish piled there at the restoration of the church in 1844, but the present rector (Rev. W. H. Langhorne) tells me that an old inhabitant says that he remembers the mound there before the restoration. On the other side of the valley is Ilbury Camp, a conspicuous object from the spot where the mound stands.

Ramsden.—To the south of Wychwood Forest on the farm road to Brize's Lodge, on ploughed land, nearly destroyed.

Sarsden.—On south of road from Churchill opposite Nursery Plantation, called 'Squire's Clump,' planted and surrounded by a wall and ditch and in good preservation, 225 ft. circumference, 13 ft. high from the ditch which is 3 ft. deep.

Another partly within Skew Plantation and partly in ploughed field, very much reduced.

Shipton-under-Wychwood.—In a plantation on Shipton Down, 200 yds. west of the spot where the road from Charlbury crosses that from Shipton to Burford, is a round barrow admirably preserved and which differs from any other specimen in this county in being partially surrounded by a rectangular earthwork. The mound is 280 ft. circumference, 10 ft. 6 ins. high. It stands on the open side of a rectangular area, the other three sides, the north-west, south-west, and south-east, being enclosed by a small bank 3 ft. high, the first-named of which is 50 ft. from the barrow and the other two 100 ft. The ground is level and there would appear to be slight traces of a very shallow ditch on the inside of the bank.

Spelsbury.—Two in a ploughed field east of Spelsbury Down Farm, almost destroyed.

Stoke Lyne.—On the south of the road from Tusmore to Souldern, planted and well preserved, 191 ft. circumference, 5 ft. high.

Swalcliffe.—On rising ground beyond the stream, a quarter of a mile due south of Madmarston Camp, is 'Rowbarrow.' It was opened about 1854, and bones, burnt wood, and sand found, but appeared to have been previously opened. It was then described as 100 ft. circumference and 12 or 14 ft. high, (fn. 23) but is now practically destroyed.

Swinbrook.—In ploughed field north of Southlawn Cottages, partially destroyed.

Another west of Pain's Farm, planted and well preserved, 200 ft. circumference, 8 ft. high.

Tadmarton.—A pair of barrows on Heath, 360 yds. north-west of camp. Totally destroyed.

Wootton.—About half a mile south of Glympton on the west of the road to Wootton, on the top of a hill, is a mound called the 'Copping Knoll,' 175 ft. in circumference, 3 ft. high.

Wychwood.—On the hill on which the moated enclosure High Lodge stands, about 300 yds. to the north-west of the Lodge, are two small barrows, on ploughed land and partially destroyed, the western one being almost levelled.

Another just inside the forest wall and 60 yds. south of Waterman's Lodge, 200 ft. circumference, and 11 ft. high from the lowest part of the ground, which slopes away here to the east.

Elongated barrow

Crawley.—In the corner of the field next to Maggott's Grove, on the top of the hill south of Crawley Bridge, was an elongated barrow, which when measured by Akerman about 1858, (fn. 24) when nearly half had already been removed, was 107 ft. by 83 ft. Since then the demolition has been practically completed, but its site is still to be traced on the high ground in the southwest corner of the field next to the plantation. It had been opened for stone, was again opened by Akerman, and again in 1864 by Dr. Thurnam and Professor Rolleston. (fn. 25) It was found to be full of skeletons, ranging west and east with feet to the latter, the skulls being very brachycephalous. Two bronze buckles and one of iron and a small bone camula were found and some shards of Romano-British pottery. Dr. Thurnam agreed with Akerman that the barrow was of the Romano-British period and belonged to a class of elongated grave mounds which might be taken for a primeval long barrow, from which, however, it differed in not varying in width and height. They are sometimes met with outside camps occupied by the Romans.

Other destroyed barrows

In addition to those barrows referred to above as being destroyed there are records of others which formerly existed.

Warton in his History of Kiddington (1782) says that there were perishing fragments of a few barrows on Shotover Hill, and Parker (Early History of Oxford) says that many mounds were formerly visible on Bullingdon Green, from one of which was produced early pottery, human bones, and burnt fragments.

Beesley in the History of Banbury (1841) mentions a barrow called 'Round Hill' on the north side of the lane leading from Bloxham to Milton, 72 ft. in length and 12 ft. high. A note in his MS. additions states that this was partially destroyed in 1867 and a skull found. He also mentions a single barrow and a pair of twin barrows on the crest of the hill on the road leading down to Brailes. They were nearly levelled at that time. In his additions he notes the destruction of a pair of twin barrows at Berryfields Farm, Great Rollright, in 1842, previously to which it was opened by him, and bones, ashes, and white sand were found. He also records the destruction in the same year of a barrow called 'Round Hill' on Portmeadow, Oxford. It had been previously opened.

In Fritwell parish, but to the south of Souldern Village and near the courses of the Portway and the Avesditch, stood a barrow called 'Ploughley Hill,' which Beesley (1841) records as then recently destroyed.

The name 'Lowbarrow House' of a house, three-quarters of a mile south of Leafield, suggests the existence at one time of a barrow at this spot.

Mr. Conder, in the paper to the Society of Antiquaries on Lyneham Barrow in 1895, records that he observed to the north of the barrow five or six low circular mounds which he thought might be the remains of barrows.

Reference table.

Defensive Earthworks, Dykes and Banks, Tumuli and Barrows, &c.

The points of the compass are given approximately, but will be found sufficient to indicate the position of the parish. Distances are to the church or centre of the village or town.

'Des.' signifies the former existence of a work now destroyed. T = Tumuli and Barrows.