The camp is further divided into two nearly equal halves by a low rampart and ditch running across the middle of it in a direction nearly parallel to the south face. The ditch of this small rampart, not more than 3 feet deep originally, and 4 feet wide at the bottom, is on the south side of the rampart, showing that it was the northern, or in point of altitude, the lower half of the camp which it was intended to divide off and defend from the higher or southern portion, after the latter had been gained. The opening through this small cross rampart was in the centre, and running from it down the centre of the camp for a distance of 66 yards in a northerly direction might be seen a depression described as the “ Mid Street” in the Plan, Plate CXLIV. It was about 10 inches deep and 10 feet wide on the surface of the turf, marking probably the site of the principal street of the camp. The central dividing rampart was not straight but curved out in the centre towards the south, thereby embracing to some extent the crown of the hill, and showing that in all probability this inner rampart no less than the outer one was thrown up with a view to defence, and intended to hold the northern or lower end of the camp after the upper and southern end had been taken. A similar division, by means of a small inner rampart, is seen in Maiden Castle, near Dorchester, and some other camps in the neighbourhood, but in this case the dividing off of the lower and consequently less commanding portion of the camp as a citadel from the upper and stronger half, seems to imply that other considerations in addition to those of defence, influenced the particular arrangement of this camp. We may observe that the upper half of this camp, which we must call the outer portion, was contiguous to the flat downland plateau on the outside of the camp which was suitable for the grazing of cattle, whereas the lower and southern portion being bounded by steep slopes, terminating probably at that time in wooded bottoms, was difficult of access for sheep and oxen. It seems likely therefore that the garrison inhabited this northern half of the camp and that the dividing rampart separated the inhabited portion of the camp from that which, being contiguous to the flat plain outside, was devoted to the cattle when driven in, and this may also explain the use of the three large openings in the outer rampart on this side which may have been so constructed for the purpose of driving quickly in, when attacked by a neighbouring tribe, the animals which in ordinary times were kept grazing on the downland to the south of the camp. This downland was probably always suitable for grazing purposes. The forest district of Cranborne Chase, which spreads over the country for some miles to the south, east, and west, does not rise higher than about 650 feet above the sealevel, that is, to a distance of about 1 mile 730 yards from the camp leaving an open downland space of about a mile in width and 10 miles in length along the tops of the Wiltshire hills contiguous to the camp, which could never, owing the slight depth of soil, have been covered with trees. Much of it was cultivated in prehistoric times, and traces of terrace cultivation and square enclosures may still be seen on the south side of the hills, but much must have remained in pasture, and herein consisted the wTealth, if wealth it can be called, of the prehistoric inhabitants of Winkelbury. Here also was the point of attraction for neighbouring marauders. An enemy coming from the north, though no doubt he would be partly concealed at that time by the wooded valley of Broad Chalke and the low district about Donhead and Ferne, would be exposed to view from the hill, and would have to ascend a steep open slope before he reached the camp, affording time for the defenders to retreat behind their ramparts. But from the recesses of Cranborne Chase on the south, an attacking party might come upon the flocks unawares. The first sign of their approach would be the signal for the shepherds to drive their cattle into the camp. From the south, east and west the cattle would converge towards the camp, and in their retreat would congregate in large numbers upon the plateau immediately outside the camp, the marauders carrying off from the outskirts of the retreating columns, such animals as remained behind. To get their cattle quickly within the entrenchment would therefore be a matter of forethought and arrangement with the inhabitants of Winkel bury. To this end, the large openings which I have described appear to have been left in the ramparts so as to enable the flocks and herds to be driven in with a broad front and this was provided for even at the expense of weakening the defences on this side. Once within the ramparts the object of the assailants would be defeated ; the outer rampart would be manned and the inner dividing rampart would be held by a reserve body and would afford security against a coup de main, and it is hardly probable, judging from the usual habits of people in a pastoral stage of civilization, that they would proceed to besiege the camp.

In front of the centre of the east face of the camp and just; outside the edge of the ditch, there was a pit surrounded by a bank, and from the pit two small tracks led, one to the east terminating on the brow of the hill, and one to the north also terminating on the brow of the hill on that side. Besides the three larger openings on the south side of the camp and the opening through the centre of the inner dividing rampart which I have spoken of, there is another entrance to the camp on the north side above the village of Berwick St. John, and from it an ancient roadway is seen running obliquely down the side of the hill to the south-west, viz., the direction of the nearest spring. There are also to be seen other trackways running terracelike down the steep hill which bounds the west side of the camp in the direction of the spring and which seem to me to imply that constant traffic was at one time going on outside and past the camp. On the east and west sides of the hill are two very deep gulleys occupied by roads running north and south, one on the east running to the lower Bridmore valley and the other occupied by the present road from Rushmore to the village of Berwick St. John. These gullies appear to have been formed in the course of years by rain-water running along the channels formed, in the first instance by mule or horse-tracks and gradually washing it down, they afford good examples of the great amount of denudation that is produced by running water in this way. That on the east is now as much as 23 feet deep and 63 feet wide on the top. That on the west side also shows evidence of continuous deepening and change of track. The old line of track after having worked down to a depth of 8 feet has been abandoned for a parallel side track which has now worked down to a still lower level of 10 feet, leaving between the two tracks an apparently high bank which local people sometimes suppose to be a rampart and to have formed part of the defences of the hill, but its summit is only on the natural level of the surface of the ground and the parallel ditches on each side of it have been excavated by water and traffic and the materials have been washed away by running water and lost in the streams below.

The nearest supply of water is now reached only at the foot of the hill to the west, at a distance of 850 yards from the camp. Whether or not this was the source of supply for the camp at the time of its occupation it is difficult to say. It is a difficulty always or frequently met with in these camps, and one which has even led some persons to assume in the face of distinct evidence to the contrary, that they could not have been camps but places of temporary resort for religious or other purposes. In this camp, however, as in many others, we shall see that there is evidence of continuous occupation. Three explanations only can be given to account for this absence of water in camps, one, that the inhabitants fetched their water from the outside, a matter of difficulty if the camp was continuously and systematically besieged, but of such a custom in early times we have no evidence, nor, I may say, viewing the habits of uncivilized people generally, is it probable that such was the case. It is probable that warfare in those days consisted of raids between neighbouring tribes and that the defenders carried in with them such a supply of water as they might require during a short attack never exceeding a day or two. The second explanation is that wells may have existed which have been filled up and have never since been discovered. The third explanation is that the springs were higher in prehistoric times and nearer the camps. We know that m many chalk districts and in this neighbourhood in particular, there are high springs which run only in the winter when the hills have sopped up the winter rains and retained it like sponges at higher levels, and in ancient times, when the valleys and hill -slopes were covered with forests, even if no other physical causes tended to produce moisture, the hills at all seasons must have been full of water shooting out in higher places than is the case at present, but of which evidence may still be seen in many districts and in many spots where no water has run within the memory of man. In the case of Winkelbury, however, it does not appear probable from the form of the hill that any spring could have run out from the hill itself.

Opinions have been hazarded by several writers as to the date and object of this camp, but I am not aware that any exploration of it has been attempted. Sir Richard Colt Hoare,^a after describing the three entrances which I have mentioned, observes that it differs materially from any camp that he has seen. He says that it contains 12½ acres, and that its circumference is 1,056 yards, which is about correct, 1,040 yards being the measurement taken by me along the crest of the rampart. He says the ramparts are 39 feet deep, but this is an exaggeration, as they now where approach that depth. His description is accompanied by a very rough plan. The Rev. F. Warne^b also mentions, and gives a rough plan of, this camp, including it amongst those which he believes to be constructed for permanent occupation rather than for military purposes. He suggests that the interior division which I have described denotes a separation for cattle, and that wherever this is wanting (and the defences simply surround the hill without any second line of defence), such camps are thrown up for temporary and warlike purposes. I hardly think that we have sufficient evidence as yet to warrant this nevertheless very reasonable supposition. Permanent occupation can be determined only by excavations, but in this instance I think it will be seen that the conjecture is very probably correct. Murray’s^c Handbook for Wiltshire speaks of Winkelbury as Vespasian’s camp, and the Bev. John Austen, in the Archaeological Journal^d, says that it is supposed to have been occupied by Vespasian, but I am not aware that there has been any reliable foundation for this assumption. It is certainly not of Roman construction.

Note a. The Ancient History of South Wiltshire , folio, 1812, page 247.

Note b. Somersetshire Archaeological and Natural History Society , Proceedings, Yol. VIII., part 2, p. 63.

Note c. Handbook for Wiltshire.

Note d. Journal of the Archceological Institute , Yol. XXIV., page 168.

WINKELBUBY CAMP.

Section I. — A section 10 feet wide was cut through the rampart at A 8 on the plan, Plate CXLIV. (see also Section I., Plate CXLV.), close to the west of the central entrance to the camp on the south face. The crest of the rampart was 6.82 feet above the old surface line, which latter was very clearly marked by a bluish seam 6 inches in thickness running under it. Nothing was found until we came to the part beneath the crest where 445 flint flakes were found lying on the old surface line, being contained within an oblong space of 4 feet 6 inches by 2 feet, as shown in the section, Plate CXLV. With these were five fragments of pottery of two qualities, one containing small grains of flint, of the first quality named in the Belie Table, page 253 ; and the other five, of the second quality — smooth, without grains, all apparently of British manufacture. In the middle of these flints was found a flint flaker, Fig. 23, Plate CXLVII., of the usual well-known form as described in Mr. Evans’ work on Flint Implements (see Fig. 349, page 370). The flakes had for the most part bulbs of percussion and facets, and were undoubtedly prehistoric flakes, there being no possibility of mistaking them for flints fractured for road making. But they were not well-formed flakes such as might have been intended for use as knives ; they were wasters or else flakes of the kind that might have been used in a tribulum^e for threshing corn. Indeed the oblong space in which they were found almost suggests tlie possibility of their having formed part of a tribulum which lay on the surface when the rampart was thrown over it. There was, however, no trace of wood, and the flakes in many cases were found touching one another. A little further on towards the ditch, also on the old surface line, beneath the rampart and close to the flakes, was found a portion of a ring of Kimmeridge shale. Fig. 14., Plate CXLVII. The presence of this characteristic Dorsetshire specimen of antiquity affords some grounds for estimating the date of the entrenchment. There can be no doubt that both it and the flakes lay on the surface of the ground when the rampart was formed. There is no possibility of so small and light an object as this shale ring having worked its way down from the present surface a depth of 4 '5 feet, nor could the flakes have accumulated on the old surface line from the top. If it could be shown that Kimmeridge shale was not used in the turning lathe before Roman times it would prove the rampart to be post-Roman ; but this is not the case. Undoubtedly the so-called Kimmeridge coal-money has been found more frequently than otherwise in association with Roman remains. It recurs in association with Samian pottery near Kimmeridge, and it was so found both by Mr. Austen and myself in the RomanoBritish Village on Woodcuts Common, which is a short distance from Winkelbury. But, on the other hand, the discovery of a lathe-turned vase of this material in a tumulus on Broaddown, Farway, near Honiton, and the excellent paper by the Rev, Richard Kirwan, in the Norwich volume of the International Prehistoric Congress, affords good grounds for believing that it was used in the lathe at a much earlier period.4 In regard to the flakes found on the old surface line, it is just possible that they may have laid on the surface in this spot since the neolithic age. Flakes chipped off and buried in long grass or boggy ground, might remain undisturbed for a considerable period, but that is not the condition of things that can ever have existed on this spot. The soil of the down is covered by a carpet-like expanse of fine turf, and the smoothness of the old surface line when it was laid bare beneath the rampart, showed that the same verdure probably existed at the time of the construction of the rampart. Unbroken by trees or shrubs, it is most improbable that so large a deposit of flakes as 455, within a space of 4 feet 6 inches by 2 feet, should not have been dispersed if they had been there for any length of time. It was noticed that they all lay on the top of the dark seam denoting the old turf and must have been covered over by the mould not long after they were deposited, at the time the rampart was thrown over them. They were in all probability flaked off at the time the camp was laid out and covered over, together with the flaking tool found amongst them almost immediately afterwards.

Note e. Two of these still in use in Cyprus and in Assyria, the survivals of similar ones used formerly by the Romans, are in my Museum in the village of Farnham, Dorset.

Note f. For the association of Kimmeridge shale with Roman remains, see Rev. H. Austen, Journ. Arch. Inst., X., p. 362 ; Hon. H. C. Neville, Journ. Arch. Inst., XIV., p. 85 ; ditto, XV., p. 84; Mr. John Sydenham, Journ. Arch. Inst., I., p. 347 ; Rev. John Austen, Journ. Arch. Inst., XXIV., p. 169 ; Sir AV. Tite, Proc. Soc. Antiq., Lond., V. (2nd Series), p. 30 ; Rev. Preb. Scarth, Proc. Somerset. Arch, and Nat. Hist. Soc., XXIII., p. 8; Mr. Roach Smith, Coll. Antiqua, III., p. 33; Rev. J. Austen, Papers of the Purbeck Society, 1857, p. 93. For its connection with flint and British remains see Warne’s Ancient Dorset, p. 19 ; Journ. Arch. Assoc., II., p. 234; Rev. John Austen, Journ. Arch. Assoc., XVI., p. 300; Mr. W. Long, Wiltshire Arch. Mag., XVI., p. 182; Rev. R. Kirwan, Trans. Inter. Prehistoric Congress , Norwich, 1868, p. 363; Mr. P. 0. Hutchinson, Proc. Soc. Antiq., Lond., IV., p. 161.

My attention was next called to the trace of a street marked mid street on the plan, Plate CXLIV., running for a space of 199 feet from the southern entrance. The whole length of it was excavated. An average section of this is shown in "Section across Mid Street," Plate CXLV. It was found that an undisturbed chalk bottom was reached at an average depth of 18 inches. Of this the upper 8 inches was turf mould without stone, and the remaining 10 inches consisted of a mixture of chalk rubble and mould. The street was originally 9 feet wide, increasing to 10 feet in places. It had no paving or stones, and the sides appear to have sloped up at a steep angle. At 199 feet from the entrance northwards, it died out gradually. On entering the gateway, the main part of the traffic appears to have gone down this street, which may possibly have been lined with huts, though no trace of them could be found ; but part of the traffic turned sharply to the eastward immediately after passing the gate. Nothing was found in the turf mould, but beneath it the chalk rubble was strewn with fragments of pottery and other debris. These consisted of 17 fragments of burnt clay, perhaps part of the daubing of wattling work, but no impressions of the sticks could be found on them ; an irregular fragment of bronze 0‘7 inch across, Fig. 10, Plate CXLVII. One rudely-formed bone pin 3.2 inches in length, Fig. 34, Plate CXLVII., one fragment of iron, perhaps part of a knife or hoop iron, one fragment of iron apparently part of a knife, Fig. 18, Plate CXLVII., one piece of square rod iron 2 '42 inch in length, Fig. 21, Plate CXLVII., one ring of iron 0.8 inch exterior diameter and 0.14 inch thick, Fig. 15, Plate CXLVII., two other small nondescript fragments of iron and several iron clinkers, several fragments of sandstone ground smooth on one side, either burnishers or used for grinding corn. Only 17 irregularly-formed flint flakes, with bulbs of percussion, were found throughout the whole length of the street, and large quantities of animal bones, a description of which is given in the Animal Remains column of the relic table. The pottery consisted chiefly of small fragments, mostly hand-made, some smooth and some mixed with grains of flint, and pieces of it were ornamented with the impress of a human nail and fingers, of which Figs. 12 and 13, Plate CXLVIII., are specimens. A few fragments, however, were lathe-turned, and apparently of Romano-British construction, of which Fig. 1, Plate CXLVIII., is a piece of a rim. One fragment was ornamented with raised bands. One fragment of Roman red glazed Samian was found at the lower end of the street. Here also we came for the first time upon a kind of pottery, which, so far as I know, is peculiar to this camp, only one specimen of it having been found at Woodcuts. Its peculiarity consists in being mixed with oolite grains. It is recorded in the 3rd column of pottery in the relic table, and is so remarkable as to afford evidence of identity of age in the various places in which it is found. It was thought at one time that the thinner and finer kinds of pottery were distributed nearer the surface than the thicker, but on further observation this idea had to be abandoned. The fine and the coarse were mixed together indiscriminately and appear to have belonged to the same period. The evidence afforded by this cutting showed very slight trace of Roman occupation ; the majority of the relics were pre-Roman in character but of the Iron Age.

Immediately below the eastern rampart outside the ditch a basin-shaped pit, about 24 feet in diameter and 18 deep, was seen surrounded by a bank on the lower side, see the spot marked Pit Dwelling, Plate CXLIV., and also Plate CXLVI.

This we excavated and found it had been a pit dwelling 12' 6" X 11' 3" feet square at the bottom, with an entrance to it on the lower or eastern side ; this entrance was 1 foot wide at the bottom, and had a wooden step consisting of a plank 1½ inches thick, the blackened remains of which were found in the silting near the consisted in the centre of 0 7 feet turf mould thinning out towards the sides, then 1.9 feet of black earth, below which was 3.0 feet of chalk rubble. The relics in the silting consisted of a piece of iron band 2 inches in length, 0.3 inch in width, and 0T4 inch thick, at a depth of two feet beneath the surface at the bottom of the black earth, two flint flakes with bulbs of percussion in the black earth, and sixteen fragments of red burnt clay, perhaps the daubing of wattle work. In the chalk rubble beneath we found a fragment of sandstone hollowed by rubbing, a fragment of a bronze pin, and several Helix aspersa. This snail does not live on the hill now.

The pottery in the black earth did not differ from that in the chalk rubble, and consisted of fragments with small grains of flint, and pieces without grains and smooth, all apparently of British manufacture.

As this was the only Pit Dwelling found on the hill, the habitations within the Camp, and to the south of it, having been probably constructed above ground, it seems questionable whether it is of the age of the Camp or earlier. It is possible it may have been upon the ground earlier, and perhaps was partly filled up from the materials excavated from the ditch, or it may have been inhabited by people who for some reason or other, were excluded from the Camp.

Section III. — I now decided to cut a section 10 feet wide through the ditch and rampart on the north side near the north entrance, in order to ascertain, if possible, whether the fragments of pottery and other relics that might be found in it, were of a nature to prove it to be of the same or a different date from those found in the pits and other places in the interior, having in view, of course, the probability that the camp was occupied in times subsequent to the date of its construction. The previous sections cut in the south face and in the cross rampart had been unsatisfactory, and the relics found in them had been too scanty to rely upon.

The columns in the relic table of this section refer to the different parts of the ditch and rampart. The necessity for considering the various parts separately is a point that I have dwelt upon repeatedly in my accounts of previous excavations, and need not be again insisted upon in this place. The denudation of the rampart and silting of the ditch and interior slope in the course of ages is always so considerable that the relics found in these different parts often represent different periods, and ought carefully to be distinguished. The results of the cutting may briefly be summarised as follows : — The pottery is of the same qualities as that found in the pits and in the Mid Street, and fell under precisely the same classification including the fragments with oolitic grains, which are such a remarkable feature in the pottery of this camp, 25 fragments of which were found quite down on the old surface line under the rampart. Of the hard black quality only two fragments were found in the body of the rampart, but 19 fragments of it were found in the silting of the ditch. It is of a superior quality to the others, but not Romano-British. The predominating quality throughout the section consisted of the second quality as was the case in the interior. Some small fragments of bronze were found on the old surface line beneath the rampart, and also some fragments of iron slag, but nothing that could be identified as a work of art of any period.

I now turned to the remains that were observed on the outside and to the south of the camp, to which I was attracted by several tumuli shown on the Plan of Winkelbury, Plate CXLIV., and some of which are shown on a larger scale on Plate CXLIX. But before describing the tumuli it may be well to speak of several small pits similar in form to those of the interior which were interspersed between and about the barrows, and are marked I. to VIII. on the Plan, Plate CXLIV. They were circular and about the same size as those in the interior, and, like them, were only noticed after some hesitation, in consequence of very slight depressions on the surface. Some were filled entirely with black mould, and others with mould half way down and chalk rubble beneath it. Amongst the contents were included one fragment of clay, evidently the daubing of wattle work, a rude chalk spindlewhorl, rubbed sandstones similar to those found in the camp, and an iron knife. One of the pits contained a small British urn and a patera of coarse material handmade, Figs. 8 and 9, Plate CXLVIII. Three of them contained a number of the perforated chalk weights similar to those found in the pits in the interior, and which I identified with those I found in pits in Mount Caburn Camp, near Lewes. Numerous flint flakes, and a portion of a bronze ring brooch. No. 5 contained a fragment of Romano -British pottery quite different from the rest, the only piece found. Its depth beneath the surface is not recorded. Two contained deposits of carbonized wheat, and one of them several sandstones, which had evidently been used for pounding wheat. The grains were very small, containing 668 grains to the cubic inch, much smaller than those found in pits at Rotherley and Woodcuts, a circumstance to which I have alluded in the first volume, page 177, where a table of the size of the grains of the different periods is given and reproduced in this volume ; from which it is conjectured that the agricultural status of these British people was inferior to that of the Romanised Britons, as might be expected. The pottery was of the same quality as found in the camp, the second quality predominating, but it is noteworthy that no specimen of the third quality containing the oolitic grains was found in these pits outside the camp. No fragment of Samian was found. Only two fragments of hard pottery of the fourth quality were noted, one of which was the piece identified as Romano-British, but as it is not recorded at what depth it was found, it may no doubt have been found near the surface. Bones of the pig and sheep were found in Pit YI. Altogether the relics seem to indicate that they were refuse pits, and no doubt more would be discovered if the ground were thoroughly trenched over, but not being on my own property and being unwilling to create any possible difficulty between the owner and the tenant, my examination of this hill was not as thorough as it would otherwise have been. The date, like that of the pits in the interior of the camp, is open to question, but the resemblance of the relics to those of Mount Caburn leads me to think that they are pre-Roman. The knife found in Pit 7 might certainly be Roman, but the inhabitants of the southern part of Britain traded habitually with the Continent before the arrival of the Romans, and the majority of the fragments of pottery are certainly British.

DESCRIPTION OF PLATE CXLVII. RELICS FROM WINKELBURY CAMP AND WINKELBURY HILL.

Fig. 1. — Iron knife, greatest length including tang 4’00 inches, greatest width 076 inch, thickness of back 0.14 inch. Found at a depth of 2 feet 9 inches in the filling of Pit 7, Winkelbury Hill, associated with a quantity of carbonized grains of wheat.

Fig. 2. — Iron knife, greatest length including tang 375 inches, greatest width 0‘95 inch, thickness at back of blade 072 inch. Found at a depth of 1 foot 10 inches in the ditch, at the junction of the Inner and Outer Ramparts on the East side, Winkelbury Camp.

Fig. 3.— Bronze awl, length 2’0 inches, greatest thickness 0.12 inch, flattened at one end. Found at a depth of 1 foot 4 inches in the Ditch of Barrow I., Winkelbury Hill, on the East side. See Section IV., Plate CXLIX.

Figs. 4 and 5. — Two thin fragments of bronze, one piece having a hole 0.06 inch in diameter. Found on the old surface line beneath the Rampart, East Extension, Section III., Winkelbury Camp. See Plate CXLV.

Fig. 6. — Fragment of bronze found whilst filling in the Pit Dwelling, Winkelbury Camp.

Figs. 7 and 8. — Two small studs of bronze, about 078 inch in diameter, found at a depth of 2 -05 feet beneath the surface on the old surface line, below the Rampart, Section III., Winkelbury Camp.

Fig. 9. — Small semicircular fragment of bronze, 0-32 inch in diameter. Found on the old surface line beneath the Rampart, Section III., Winkelbury Camp.

Fig. 10. — Irregular fragment of bronze, 070 inch across, obtained from a depth of 12 inches in the Mid-Street, Winkelbury Camp.

Fig. 11. — Portion of bronze ring brooch, ornamented with incised “ hatching”; found at a depth of 2 feet 2 inches in the filling of Pit 6, Winkelbury Hill.

Fig. 12. — Flint scraper, found at a depth of 14 inches beneath the surface in the ditch of Barrow I., Winkelbury Hill.

Fig. 13. — Flint scraper, found at a depth of 370 feet beneath the surface in Barrow I., Winkelbury Hill. See b, Section I., Plate CXLIX.

Fig. 14.— Half a ring of Kimmeridge shale, 1'20 inch exterior diameter, 0.22 inch in width, 0-22 inch thick. Found 4 '5 feet beneath crest of rampart on the old surface line, Section I., Winkelbury Camp.

Fig. 15. — Portion of iron ring, 0.80 inch exterior diameter, 0T4 inch thick. Found in the Mid-Street, Winkelbury Camp.

Fig. 16. — Disc of baked clay, 1 '6 6 inch diameter, 0.68 inch thick, with two dents, one on each side 0-26 inch in diameter. Found in the Jilling of Pit 6, Winkelbury Hill.

Fig. 17. — Fragment of iron, perhaps the point of a knife, l’l 6 inch long, 0T0 inch thick. Found at a depth of 8 inches in the Mid-Street, Winkelbury Camp.

Fig. 18. — Fragment of iron, perhaps part of a knife or hoop-iron, P56 inch in length, 0T0 inch thick, 0.52 inch in width. Found in the Mid-Street, Winkelbury Camp.

Fig. 19. — Fragment of iron 0’06 inch thick, P62 inch long and 0.50 inch wide, consisting of two pieces apparently intended to enclose a piece of leather. Found in the filling of Pit 5, Winkelbury Camp.

Fig. 20. — Fragment of band iron 2'00 inches in length, 0’30 inch in width, 0T4 inch thick. Found 2 feet beneath the surface in the centre of the Pit Dwelling.

Fig. 21. — Piece of square rod iron, 2 *42 inch in length, 0.22 inch square. Found in the Mid-Street, Winkelbury Camp.

Fig. 22. — Rude chalk spindlewhorl 1'50 inch in diameter, 1 inch thick, perforated with hole 0.40 inch diameter. Found in the filling of Pit 1, Winkelbury Hill.

Fig. 23. — Flint flaker, chipped all round and worn at the end, 2 ’50 inches long, 0’50 inch thick. Found with a cluster of 455 flint flakes, on the old surface line at a depth of 6 ’8 6 feet beneath the crest of the Rampart in Section I., Winkelbury Camp. See Section I., A, Plate CXLV.

Fig. 24. — Flint knife, chipped all round. Found in Barrow I., Winkelbury Hill.

Fig. 25. — Chalk weight, 6 inches long, 4 inches wide, and 2 inches thick, similar to those found in pits at Mount Caburn and Cissbury, near Worthing. Found, with 18 more, in the filling of Pit 7, Winkelbury Hill.

Figs. 26 and 27. — Two flint scrapers, found on the bottom of Pit 5, Winkelbury Camp.

Fig. 28. — Flint scraper, found in Hollow Depression No. 4, Winkelbury Hill.

Fig. 29. — Flint scraper, found in a circular cutting south of Pit 5, Winkelbury Camp.

Fig. 30. — Implement of stag’s-horn, 5 inches long, found in the filling of Pit 6, Winkelbury Camp.

Fig. 31. — Piece of stag’s-horn, 2 '4 inches in length, not perforated. Found in the filling of Pit 6, Winkelbury Hill.

Fig. 32. — Fragment of bone comb, ornamented with wave-bands, filled with dots and apparently burnt. Found in the filling of Pit 1, Winkelbury Camp.

Fig. 33. — Piece of sandstone with one flat surface, and showing marks of attrition. Found in the filling of Pit 8, Winkelbury Hill.

Fig. 34. — Rudely -formed bone pin, 3 ‘20 inches long. Found in the Mid-Street.

Fig. 35. — Fragment of the daubing of wattle work. Found in the filling of Pit 1, Winkelbury Hill.

Fig. 36. — Tip of cow’s-horn, T94 inch in length, perforated as a handle. Found at a depth of 2 feet 1 inch in Pit 5, Winkelbury Hill.

Fig. 37. — Bone gouge, 5 inches long. Found on the bottom of Pit 5, Winkelbury Camp.

Fig. 38. — Piece of sandstone with one flat surface and showing marks of attrition. Found in the body of the Rampart, Section III., Winkelbury Camp.

DESCRIPTION OF PLATE CXLVIII. URNS AND POTTERY FROM WINKELBURY CAMP AND WINKELBURY HILL.

Fig. 1. — Fragment of hard pottery of sandy texture ornamented with two raised bands 0 50 inches apart. Found in the Mid-Street of the camp.

Fig. 2. — Urn of brown, hand-made pottery, ornamented with chevrons filled with lines. Height, 4\ inches, diameter of mouth, 3 inches. Found at a depth of 13 inches beneath the surface in the centre of Barrow No. IY. on Winkelbury Hill. The urn was filled and surrounded by minute chips of flint. No burnt bones were found with it.

Figs. 3 and 5. — Two fragments of thin, smooth, brown pottery without grains, 0T8 inch in thickness, and ornamented with an incised zigzag pattern. Found in Pit 4, Winkelbury Hill.

Fig. 4. — Fragment of brown, sandy pottery, 0.32 inches thick, ornamented with a raised band 0T2 inches wide. Found at a depth of 2 '35 feet beneath the surface in the ditch of the Rampart, Section III., Winkelbury Camp.

Fig. 6. — Fragment of rim of smooth red pottery, 0.24 inches thick, ornamented with two and three rows of punch marks. Found in the filling of Pit 8, Winkelbury Hill.

Fig. 7. — Fragment of red Samian pottery, 0T2 inch thick, with remains of raised ornamentation. Found at a depth of 7 inches beneath the surface in Hollow Depression, No. 3, Winkelbury Hill.

Fig. 8. — Hand-made cup of greyish pottery without grains, 2'2 inches high. Found at a depth of 2 feet 8 inches beneath the surface on the bottom of Pit 1, Winkelbury Hill.

Fig. 9. — Hand-made vessel of reddish-brown pottery, 3 ’4 inches high, 3 '4 inches greatest diameter, and 3 '14 inches diameter at rim. Found on its side 2 feet 8 inches beneath the surface at the bottom of Pit 1, Winkelbury Hill.

Fig. 10. — Fragment of rim of sandy pottery, 0.60 inch wide, 0.28 inch thick. Found in the Mid-Street, Winkelbury Camp. This class of rim has been found associated with Samian in the Romano-British villages at Woodcuts and Rotherley.

Fig. 11. — Fragment of smooth brown pottery, 0‘36 inch, ornamented with 3 punch marks. Found in the filling of Pit 1.

Fig. 12. — Fragment of coarse, hand-made, brown pottery, 0.32 inch thick, ornamented with a row of punch marks. Found in the Mid-Street, Winkelbury Camp. Fig. 13. — Fragment of rim of smooth hand-made brown pottery, 0*30 inch thick, ornamented with impress of the finger nail. Found in the Jilling of Pit 6, Winkelbury Camp.

Fig. 14. — Fragment of smooth brown pottery, 0-34 inch thick, ornamented with 5 punch marks. Found in the body of the rampart, East Extension, Section III., Winkelbury Camp.

The quality of the pottery found in the pits on Winkelbury Hill differs from that of more recent date, in being smooth and uneven on its surface, whilst that of the Roman Age, viz., Figs. 1 and 7, is of sandy texture and better baked.