Plate I. — Sketch-plan of Maumbury Rings, similar to that given in the former Reports, the position of the 1913 excavations (Cuttings XXX. to XXXV. inclusive and two narrow cuttings parallel to the transverse axis) having been added. It shows the relative position of the cuttings made in 1908, 1909, 1910, 1912, and 1913, but the scale is too small to attempt to show structural details. The position of the Well is seen in Cutting XXXIII., and Shafts XII. to XVII. on the E. side of the central area. It should be noted that a dotted circle, passing through the middle of the shafts, and having a diameter of 169ft., has been described on this plan.

Plate II. — General view of Maumbury Rings (September 24th), taken from the top of the Great Bank, looking S.S.E., and showing the excavations of 1913 in progress. The W. terrace is seen in the foreground, and on the opposite side the whole length of the E. terrace. The jalanks and windlass represent the position of the Well (Cutting XXXIII.). On the left the excavation of Cutting XXXI. is in progress; and further south Cutting XXX., the largest excavation made during the whole of the investigations, including the digging into the Great Bank. The photograph shows the horizontal stratification of the material forming the Civil War Terrace, the solid chalk arena wall (with strut-holes on the top), and the oblong enclosure recessed into the "wall."

Plate III., fig. A.— Cutting XXXIII., the Well, taken from the N., September 17th, 1913. The spade rests on the solid chalk arena which had been cut through subsequently for the purpose of sinking the well (4ft. in diam.). This view shows the steps, or foot-holds, cut into the S.S.E. face of the Well; there were thirteen steps on each side. The Well was apparently never completed, the bottom being reached at a depth of 27ft. below the surface of the turf. Owing to previous mutilation of the solid chalk the mouth of the Well was "steaned" with Purbeck stone slabs on the N.W.

Plate III., fig. B. — Cutting XXX., taken from the slope of the Terrace, looking S.W. (September 24th, 1913). The view clearly shows the upper margin of the mouths of Shafts XII., XIII., XIV. and XV. (counting from the S. end). On the right the solid chalk arena-floor is seen, and on the left part of the platform of the enclosure which was about l-5ft. lower than the arena-floor. The Inner and Outer Trenches are seen in section at the S. end, and in the foreground part of the Outer Trench cut into the solid chalk at the foot of the "wall."

Plate IV. — Cutting XXX., on the E.S.E. side of the Rings, taken from the N.W. on October 2nd, 1913, at the close of the excavations, and after the whole of the solid chalk in the cutting had been laid bare. The upper figure stands on the solid chalk below the Great Bank; and the lower figure on the platform of the enclosure recessed into the arena-wall. Along both sides of this area post-holes are seen, and at the S. end a recess in the wall. Strut-holes can be traced on the top of the wall on both sides of the photograph, and in the foreground the extreme E. margin of the line of shafts. The old turf line under the Great Bank is clearly defined; also the oblique seams of rubble forming the earthwork, and the horizontal stratification of the Civil War Terrace.

Plate V. — Cutting XXXIL, outside the Great Bank on the N.N.W., October 3rd, 1913. This view, taken from the N.N.W., shows the stratification of the seams forming the earthwork. The old turf line, at a max. depth of 15ft. below the crest, is clearly defined, and below it the natural solid chalk is seen. The large lumps of chalk in the foreground (left-hand side) were obtained from the earthwork. At the foot of the bank a trench, apparently of modern construction, is shown re-excavated, and the nature of its loose filling is seen on the face of the cutting.

In our last report reference was made to the unfinished exploration of the amphitheatre known as "King Arthur's Round Table" at Caerleon. It was hoped that a fund of £500 might be raised to purchase the site, complete the excavations, and put the walls of masonry into such a state of repair as to enable them to withstand the weather.

Unfortunately the Caerleon committee has been dissolved, and the scheme is in abeyance.

Beyond the city wall at Caerwent [Map] and on its N.E. side, as noted in our last report, a structure was discovered in September, 1912, which was at first thought to be a second Roman amphitheatre, but later explorations show it to be a round temple enclosing an octagonal structure.¹

In connection with the pre-history of Maumbury — its shafts, &c. — we look forward to the results of the systematic excavations which are being carried out at the Grime's Graves, Weeting, Norfolk, by the Prehistoric Society of East Anglia. The Grime's Graves consist of 254 saucer-shaped depressions which are the mouths of shafts excavated in the chalk rock. One of these shafts was excavated by Canon Greenwell in 1870, and found to be 39 feet deep with galleries at the bottom. Among the objects discovered were seventy-nine red-deer antler picks (all below 17ft. from the surface), more or less complete, a ground axe of basaltic stone, cup-shaped vessels of chalk supposed to be lamps, and a well-made chalk phallus (now in the British Museum).

Note 1. Archaeologia, LXIV., 447—452.

As director of the excavations, I had the pleasure of con- tinuing the work at Maumbury in 1913 from September 4th to October 4th (the filling-in being completed subsequently). The sub-Committee, consisting of Dr. H. Colley March, F.S.A. (Chairman), Captain J. E. Acland, F.S.A. (Secretary), Mr. J. Meade Falkner, Mr. C. S. Prideaux, and Mr.W. de C. Prideaux, rendered me most valuable support.¹ These antiquaries were frequently on the ground, and their assistance from time to time in the general organization of the investigations and overlooking of the workmen was extremely helpful to the director.² Mr. Sebastian Evans, who did much for the 1912 excavations, offered his assistance, which was readily accepted and greatly appreciated. Mr. C. S. Prideaux again rendered the Committee great service by lending his camping outfit, and although he could not be present during the whole of the operations this season, his interest in the work was as keen as ever.

Note 1. Since the last series of excavations in 1912 the Sub -Committee has lost a valued Member in the person of the Rev. C. W. Whistler, who died at Broadwey on June 10th, 1913.

Note 2. A maximum number of ten men was employed for the excavations, with John Lush as foreman.

The director has held himself responsible, as in former years, for the recording of the work, the preparation of all the plans, sectional drawings and photographs¹,as well as the care and repair of the relics discovered. Help in the matter of identifying natural history specimens has been kindly rendered by Mr. Clement Reid, F.R.S., Mr. E. T. Newton, F.R.S., and the Royal Botanic Gardens, Kew.

Note 1. Subscribers may see the full series of photographs (1908-1910, 1912, and 1913) — considerably over 100 — on applying at the Dorset County Museum.

In conjunction with this, the Fifth Interim Report, subscribers are recommended to read the previously published papers on the subject, to enable them to interpret the full significance of some of the details of structural interest; they are published in the Proceedings, Dorset Field Club, and also issued separately. The sketch-plan (Plate I.) is intended merely to show the general outline of "the Rings" and the relative position of the thirty-five cuttings which have been made.

During the season the investigation of the Well (Cutting XXXIII.), the N.W. margin of which had been found at the close of the excavations in 1912, was completed (Plate III., A). It appeared to have been sunk in the XVII. Century by the Parliamentarians, who, fearing the loss of their usual water supply, evidently decided to sink a well inside the earthwork. It was never finished, and it is probable that circumstances did not necessitate its com- pletion.

We had hoped that time would have permitted us to open up the arena-wall and the inner and outer trenches on the E. side, from the "transverse axis" to the E. end of Cutting II. Extension near the N. Entrance. But the excavation of Cutting XXX. (Plate I.) entailed such an enormous amount of labour that only a comparatively small digging (No. XXXI.) could be made adjoining Cutting II. Extension, the inter- mediate ground, left untouched, covering a maximum length of 38-75ft.

The usual structural features were revealed in Cuttings XXX. and XXXI., the former digging including the removal of the large quantity of material which filled an enclosure recessed into the arena-wall — an area of similar dimensions to the corresponding one excavated on the opposite side of the Rings in Cutting XX. (1910). A line of five shafts was also investigated in Cutting XXX. (Plate III., B). All these features will be described in their proper place.

But, perhaps, the most important work of 1913 was that carried out with a view of ascertaining the date of the Great Bank enclosing the shafts and arena. Eor this purpose the excavation of Cutting XXX. (Plate IV.) was continued towards the E.S.E. as far as the middle of the crest of the encircling earthwork, and a similar cutting (No. XXXII.) was made half-way through the Great Bank on the N.N.W. from the outside of the Rings (Plate V.). In both these cuttings the old turf line was found to be clearly defined at a depth of 15ft. below the crest of the bank, and in Cutting XXXII. it was slightly higher than the general level of the field on the N.W. at the present day. As a trench — appar- ently of late date — was found at the foot of the bank in Cutting XXXII. (Plate V.), two small excavations were made on the N. and E.N.E. (Cuttings XXXIV. and XXXV., Plate I.) to ascertain if this trench continued round the earthwork. It was clearly defined in both these places.

There is strong negative evidence that the Great Bank is of the same date as the shafts. The few relics found in the two cuttings are such as have been obtained in the shafts, and nothing which could be definitely assigned to the Bronze Age, or the Late-Celtic or Roman period, was revealed in this part of the investigations. Not a single object was obtained from the actual surface of the old turf, but in Cutting XXX. an antler pick (No. 395), of the same type as those found in the shafts, was uncovered about 1ft. above the original surface; and in Cutting XXXII. an antler rake (No. 412) was obtained within 0.075ft. of the old surface, and a burr and lower part of a red-deer antler (No. 394) only a foot above the same level. Fragments of antler were met with in two other positions in the body of the earthwork, and a piece of carved chalk of circular section (No. 409), similar to other carvings of the same character found in the shafts.

As long intervals have occurred between our excavations, and as these were filled in after each season's work, it has been no easy matter to determine whether or not the solid chalk arena was cut down to a dead level; but perfect accuracy does not appear to have been achieved. It is now found that the greatest deviation from the horizontal (as ascertained from the parts excavated) is from the S. corner of Cutting XX. to the E. end of Cutting II. Extension (Plate I.), the fall from W.S.W. to E.N.E. being 1.32ft.¹ On the other hand a level of the floor taken close to the Well (Cutting XXXIII.) on the N.N.W. agrees exactly with the arena-floor at the S. end of Cutting XXX. on the E.S.E., but some of the intermediate levels taken varied to the extent of O'SSft. Again, it was ascertained in 1908 that from the centre of the arena to the N. Entrance there was a gradual fall of 0.75ft., whereas the floor was found to be level from the centre up to the margin of the so-called "den" on the S.S.E. The lowest part of the arena appears, therefore, to have been between Cutting II. Extension and Cutting XXX.

Note 1. The stratification of the chalk rock dipped from N.N.W. to S.S.E.

In speaking of the Roman work it may also be recorded here that the enclosed platform of solid chalk in Cutting XX. on the W.N.W. was 115ft. higher than the nearest part of the arena-floor, whereas the corresponding platform in Cutting XXX. on the E.S.E. was 15ft. (average) lower than the adjacent arena. The maximum dimensions of each of the enclosed platforms were 1575ft. by lift. The so-called "den" (Cutting XV., 1909) measured 17 5ft. by 135ft., and was absolutely level with the adjacent arena- floor.

From the excavation of Cutting X. (1908) on the N. and Cutting XXX. (1913) on the E.S.E., it has been clearly shown that the vertical depth from the old turf line under the Great Bank to the adjacent arena-floor is 1175ft.; and as the average depth between the old turf line and the solid chalk was found to be 2ft. in Cuttings X., XXX., and XXXII., it is evident that a depth of 975ft. of virgin chalk was removed in the central area subsequently to Neolithic times, and presumably by the Romans.

As Plate I. shows, the position of seventeen shafts has now been determined, and a dotted line has been indicated on this plan showing that the middle of each shaft is on the line of a true circle having a diameter of 169ft. Using the same centre¹ it is found, from Mr. Feacey's plan, that the general line of the crest of the Great Bank is practically a circle with a diameter of about 276ft., and the diameter of the outer boundary of the earthwork is about 335ft.

Note 1. The centre of the 169ft. circle is at a distance of about 10ft. S.W. of the central picket in the arena used for surveying purposes.

The arena-floor, on the other hand, is a wide oval with diameters of 1925ft. and 158ft. (measured from the inner margin of the "inner trench"), and the outside dimensions of the earthwork including the bulge at the S.S.W. are 345ft. on the long axis and 335ft. transversely. Within the margin of the solid arena no trace of a shaft has been found, although several small cuttings have been made partly with the intention of testing this point.¹

Note 1. These cuttings were Nos. III., IV., V., VI., XI., XIII., XIV., XXXIII.; also two small unnumbered cuttings between Cutting XXX. and the centre (vide Plate I.).

Assuming that the prehistoric shafts existed before the solid chalk of the central area had been lowered — and the arena-floor is estimated to be 1175ft. lower than the original ground level — it appears quite evident that, at the surface, they did not originally take the form of pits at all. It is the opinion of the director that in the position of these shafts, in prehistoric times, there existed an immense circular trench, or ditch, having a medial diameter of about 169ft. Judging from the excavations in Cutting XXX. and elsewhere, this ditch was probably some 16ft. deep below the original surface, and perhaps something like 40ft. wide at the top. Estimated by the average slope of the sides of the shafts it is evident that they could not have shown themselves on the original surface as pits, and, indeed, it is seen by a glance at the plan and sections that the mouths of the majority of the shafts united below the level of the arena-floor. Had we re-excavated the long hollows, called Shafts IV. and XVII., to a greater depth, it is probable that they would have been found to divide into several pits.¹

Note 1. It is possible that these dimensions may have to be altered to some extent. The approximate width at the top is given on the assumption that the counterscarp of the fosse was at a very much steeper pitch than the escarp; but if the inclination of the escarp and counterscarp did not vary much, then there must have been a decided berme between the earthwork and the fosse.

It appears, therefore, that a large circular fosse was originally cut to obtain material to form the encircling bank, and that shafts of various shapes were sunk from the bottom of this trench. On the other hand it is possible that the shafts may have been excavated first, and before any regular bottom of the fosse was cut.

We have nothing further to add Avith regard to the purpose of the shafts since the last part of Section V. of the 1912 Report was written, beyond what is mentioned above. It is not known how long they were left open; but it is evident that they were not filled in by ramming (except just at the mouths), as the rubble was found to be very loosely compacted. There can be no doubt that more chalk was excavated from the shafts than found its way back into them. The additional material might have been used to increase the height of the earthwork.

In the former reports the depth of the shafts has been given as measured from the nearest part of the arena-turf. Now that the old surface line under the Great Bank has been exposed in three places, it is possible to give their approximate depth below the original surface of Neolithic times. On these lines it is estimated that the average depth of the seven re-excavated shafts was originally 35ft., whilst their average depth below the nearest arena-turf of the present day is exactly 27ft.¹

Note 1. The same depth as the unfinished Well. (Cutting XXXIII.).

The Roman settlers at Dorchester, feeling the need of an amphitheatre outside their walls, and finding prehistoric Maumbury in a suitable position, appear to have adapted this site to their requirements, lowering the central area to convert it into an arena and leaving the virgin chalk in situ for the core of the boundary -wall. There can be no doubt that they found the prehistoric trench open, or rather only partly silted up, and in some cases the upper part of the mouths of the shafts would still be open. This is proved by the fact that in the rammed chalk in the opening of the shafts, and below the level of the arena-floor, we have constantly found Roman remains mixed with broken antler picks and flint implements of Neolithic type. The Romans had not only to cut out their arena-floor, but to make good the surrounding ground excavated by the former race by means of rammed and puddled chalk, which in places, as would be expected and as we found, had sunk over the position of the shafts below the level of the solid chalk arena. As yet there is no proof that the Romans increased the height of the Great Bank, which has, no doubt, from natural causes, somewhat shrunk during the many centuries of its existence.

The Civil War terraces, which were placed against the prehistoric bank and on the Roman deposits, have been described elsewhere.

There are features at Maumbury which resemble Avebur3 r, and the interior fosse in both cases may have been intended for the same purpose, perhaps to prevent animals and the ordinary people from trespassing on a spot reserved for ceremonies conducted by privileged peoj)le. The forth- coming excavations at Avebury (Easter, 1914), both into the fosse and vallum, may result in strengthening a comparison between these two prehistoric enclosures.

The Well (Plates I., II., and III.).

This cutting, which measured 12ft. by 9ft., partly over- lapped a small extension of Cutting XXI. (1912). On the E., S., and W. the Roman arena-floor was reached at an average depth of 3"4ft. below the surface. The floor was more or less stained, as in other places, owing to long exposure during the use of the site as an amphitheatre. The surface was rather rough, but covered with the "shingle" previously described.

In the N. corner of the cutting, solid chalk was reached at a depth of 6.3ft. and had the appearance of being fresh cut. In the central area, on the E., S., and W., it was also found that the ground had been cut out to a greater depth than the arena-floor, a chalk "wall," 175ft. in height, connecting the Roman and more recent levels. (Plate III., A.).

On digging deeper in the centre, where loose filling existed, the mouth of a circular shaft, or well, was revealed, having a diameter of 4ft. In clearing this area a few fragments of pottery (one piece glazed) were collected, none of which was earlier than the XVII. Century.

On the N. side of the shaft some slabs of Purbeck limestone were uncovered (Plate III., A.), and they were found to continue round its margin towards the W.N.W. for a distance of 3ft. (rather more than a quarter of the circumference of the hole). It was observed that the well was "steaned" only in the position where the solid chalk had been cut to a depth of 6"3ft. The width of the steaning was about 1ft., and what remained was built in four courses, the bottom being 675ft. below the surface. The stones on the inner edge were trimmed to conform to the circular shape of the well. It was found that the shaft extended downwards with the same diameter of 4ft.

At a depth of 8ft. below the well's mouth a typical fragment of glazed stoneware of the Bellarmine type (No. 338) was found. At 125ft. a piece of red earthenware with a dark brown glaze (No. 339) was obtained; this was of the same type as the albarello (No. 295) found in the New Ditch (Cutting XXIX.) outside the N. entrance, and referable to the middle of the XVII. Century. At 14ft. three glazed shards and an iron horse-shoe nail (No. 341) were found — also modern.

At this stage in the operations we temporarily stopped the re-excavation, but ultimately decided to obtain some builder's men, a windlass and other tackle (Plate II.) to pursue the work further.

Two pieces of black pottery (No. 379) were found at a depth of 143ft. below the well's mouth. One fragment is modern; the other I am inclined to regard as Romano-British, but a single fragment of Roman pottery where the whole country teems with such shards affords no evidence of date by itself, and it might easily have become mixed with the material used for filling the well. The iron nail (No. 382), depth 19ft., has a decidedly modern appearance.

At 15ft. below the mouth of the well the pieces of chalk rubble became very large, and at 16ft. Purbeck slabs began to be found and continued to the bottom. About eleven dozen of these stone slabs were afterwards counted, and most of them were shaped, one edge being slightly concave. From this fact it was evident that at one time the mouth of the well was steaned more extensively than it was when we found it, and it is possible that the stones originally extended all round the mouth.

Nodules of flint were occasionally found in the filling down to 19ft. The chalk increased in moisture at the lower levels, but even at the bottom it was not really wet. To test the true bottom the solid chalk was cut away with a pick-axe to a thickness exceeding 6ins.

Resting on the bottom of the well, thin ironwork (No. 386) was found in a fragmentary condition. Some of this was thin sheet iron bent over to double the thickness. Three pieces resemble thin door-hinges in form; three other pieces are of stouter material; these have rivet-holes at more or less regular intervals, and some of the rivets still remain, to many of which oak adheres. In some places a coarse woven fabric is seen between the wood and iron. These remains do not appear to have formed part of a bucket.

Conspicuous features in the structure of the well were the steps, of footholds (Plate III., A.), which occurred in vertical order on opposite faces, i.e., on the N.N.W. and S.S.E., thirteen steps on each side. On the N.N.W. the lowest step was 2.2ft., and the lowest step on the other side 3"2ft., from the bottom of the well. The steps in both lines were at somewhat irregular distances apart, but the average was 155ft.; their average size was — Depth, 5ins.; height, 5ins.; width, 9ins.

The bottom was reached at a depth of 27ft. below the surface, and 21.85ft. below the solid chalk margin of the mouth of the well.

There can be little doubt that this well was sunk in the troubled times of Charles I., when Maumbury was a Parlia- mentarian fort. All the relics, with perhaps one exception, appear to be of the XVII. Century. It is highly probable that the Parliamentary forces, fearing the loss of their usual water supply, decided to sink this well, and it would appear that circumstances did not after all necessitate its com- pletion. It is not likely that water was found at so high a level, and no silt, or any other soft material, was found at the bottom.

(Plates I., II., III., and IV.)

This was by far the largest cutting which has been made at Maumbury. Its position is on the line of the transverse axis and on the E.S.E. side of the Rings. It extended from the E. margin of the arena through the highest part of the Civil War terrace, and half-way through the Great Bank to its crest (Plates I., II.). The horizontal length of the cutting, in the direction indicated, was 63ft., and the width on the curve of the arena about 47ft. The margins in other directions were irregular and governed by the structure revealed as the excavations proceeded. At this spot the highest part of the terrace is 12.7ft., and the crest of the Great Bank is 22.9ft. above the arena-turf.

The arena-floor was reached at the S.W. end of the cutting at a depth of 3.75ft. below the surface, and at the N.W. end 4.35ft. Not far from the W. margin a fine chalk rubble occurred at the floor level, and it soon became evident that shafts of a similar character to those discovered in other cuttings also existed here; and the W. half of their openings gave the edge of the solid arena-floor a sinuous outline (Plate III., B.).

Comparing Cuttings XX. and XXI. (1910 and 1912) on the W.N.W. side of the arena with Cutting XXX. (1913) on the opposite side, very few structural features were disclosed. In clearing away the many tons of material from the old surface on which the Civil War terrace had been raised, the usual number of shards and other objects (mostly of the XVII. Century) were collected, including part of a glazed tile (No. 361) and about 13J dozen bullets of lead (No. 336) found under the turf in the middle of the slope of the highest part of the terrace, spread over an area some three yards in diameter, few of them being found at a greater depth than 05ft.

Judging from the condition of these bullets (many being consider- ably flattened) they had been discharged from a gun or pistol. From those in good condition it is ascertained that they vary in diameter from 14-5mm. to 18 - 5mm. They seem to afford evidence of target practice, the shots being fired probably from the corresponding terrace on the W. side. The director having recently seen similar bullets from Naseby in Warwick Museum, asked Lieut. -Colonel A. Leetham, Curator of the Royal United Service Institution, to compare the Maumbury bullets with some in his charge, and he reports that the Dorchester specimens are similar in shape and size to the examples from the battlefields of Naseby and Marston Moor and the siege of Athlone, and he sees no reason why they should not be of the Charles I. period. "The calibre of the bullet varied considerably, as did the musket of that day; but as a matter of fact the shape of the bullet did not vary until the Brown Bess went out, and indeed the round bullet was in use up to the time of the Crimean War. Mr. ffoulkes' solution may be the correct one, as the pistols of the XVII. Century were of large and varied calibres, and there are such weapons in the Institution's Museum which would take either of the two bullets you send."

Mr. Charles ffoulkes, F.S.A., Keeper of the Tower Armouries, who has seen some of the Maumbury bullets, wrote: — "The bullets are more probably pistol bullets, as all the arquebuses that are at the Tower are of much larger calibre. In the XVII. Century the pistol was fired at point blank; some writers advised ' touching the enemy's breast- plate with the pistol before it was discharged.' If your find is of different calibre it would bear out the theory that they are for pistols, as each regiment had as near as possible the same calibre of arquebus for obvious reasons. The disadvantage of having pieces of different bore was found in Elizabeth's reign, when the ' caliver ' was intro- duced to give uniformity."

After the removal of the terrace the top of the solid chalk arena-wall was soon disclosed, also the outline of an area deeply recessed into the wall, which proved to be an enclosure of similar character and proportions to that found on the opposite side of the Rings. (Plates III., IV.). By degrees the floor of the enclosed area, bounded on three sides by chalk walls and open towards the arena, was cleared, and in carrying out this work a good number of relics and shards of the Roman period were collected. The platform was found to be smooth and well worn, but there was a decided fall from front to back amounting to 0'7ft. Owing to the presence of shafts and the consequent disturbance of the chalk rock in this position the W. margin of the platform was somewhat irregular. The dimensions, however, were exactly the same as the W. enclosure, viz., length 15.75ft., width lift.

Along both sides and at the foot of the wall of the enclosure were two trenches, that on the S. 6.75ft. and that on the N. 6ft. long; they were 1.1ft, deep below the platform and of an average width of 1.2ft. at the top (Plate IV.). At each end of the trench there were single post-holes, one D-shaped, one round, and two square. All of them were 1.85ft. deep below the platform. These trenches correspond exactly with those found in the W. enclosure. The chalk wall at the back of this area reached a height of 10ft.; the lower half stood at an angle of about 80°. There was no trench at its foot, nor was there one at the foot of the corresponding wall in the W. enclosure. A deep recess of semi-circular section penetrated the wall at the W. end of the S. trench; this recess extended to a height of 3.9ft. above the floor. Its base was 0.3ft. above the platform, whereas the bottom of the similar recess in the S. wall of the W. enclosure (Cutting XX.) was 2ft. above that platform. If the posts in the S. side- trenches carried hoardings these recesses would have been obscured.

Near the top of the arena-wall to the N. of the enclosure three and to the S. two strut-holes were noted, and they were of a similar form to those found elsewhere in previous years. (Plates II., IV.).

Owing to the existence of shafts between the arena-wall and enclosure and the nearest part of the solid arena-floor, the inner and outer trenches, which bounded the arena in all parts of the Roman amphitheatre, were very difficult to trace in this cutting, for the reason that they had been almost entirely cut in the rammed filling above the mouths of the shafts, and indeed no part of the inner trench had sides or bottom of solid chalk. Here and there the outer trench penetrated the solid chalk (see foreground, Plate III., B.), but even in those places its sides were composed of rammed chalk in the upper parts. The average width of the gangway was 3ft. (as elsewhere), and the average width occupied by both trenches was 7.2ft. No post-holes could be traced in the inner trench, but in the outer trench two circular post-holes were noted at the S. end of the cutting, a square one in the middle of the front of the enclosure, and five square ones at the N. end.

From this cutting, in conjunction with others, it appeared probable that long before the existence of the enclosure and the inner and outer barriers, there was a prehistoric trench of large proportions between the arena-wall and the solid arena-floor on the line of the shafts discovered. This cutting added five more pits to those previously known, viz., Shafts XII., XIII., XIV., XV., and XVI. (Plate III., B.). Only part of the N. margin of Shaft XII. was traced at the S. end of the cutting; of Shaft XIII.¹ the N. and S. margins were cleared (giving a width of only 3.75ft.); of Shaft XVI. at the N. end only the S. margin could be determined. But the outlines of Shafts XIV. and XV. in the central part of the cutting were entirely re-excavated.

Note 1. In the mouth of this shaft part of a pick (No. 357) and the point of an antler tine (No. 359) were found.

Besides the antler pick (No. 402), depth 9ft., and an antler rake (No. 392), depth 6.5ft. below the arena-turf, a few other fragments of antler were found at a depth of 8.5ft. Below this very little was discovered until, at a short distance from the bottom, a fine rounded nodule of flint about 5iins. in diameter was brought to light. Within a foot of the bottom a pick formed from the antler of a slain red-deer (No. 405) was discovered; the skull part was slightly charred, the bez- and trez -tines were carefully reduced to stumps, and the brow-tine bore evidence of considerable wear. Below this a little burnt matter and charcoal (too fragmentary for preservation) were observed; and a large lump of chalk (No. 408), length 7ins., scratched with deep parallel incisions, with a tapering hole penetrating one surface to a depth of 1⅛in., and having a diameter of 1⅛in. at the top.

Shaft XV. (Plate I.). — This pit was of similar form to Shaft XIV., having at the mouth a long diameter of 14ft. from E. to W. and a short diameter of 8ft. from N. to S. (The W. end of its mouth is seen in the foreground of Plate III., B.). It was found to be 26.7ft. deep below the nearest arena-turf, and 34.5ft. below the old turf-line under the Great Bank. At 6.2ft. from the bottom there was a definite constriction in the walls of the shaft with a steeper pitch to the base; the diameter at this point varied from 3.2ft. to 3.5ft. The bottom was very smooth, basin-shaped, and quite circular in section, the diameter at 1ft. from the base being only 1.7ft. Nodules of flint in the filling — tried or otherwise — were not plentiful.

Parts of the crown of two antlers (No. 370), one bearing traces of fire, were found in the mouth of the shaft, depth 6.2ft. (the following depths are below nearest arena-turf); and at a depth of 7.6ft. portion of the antler of a slain red-deer (No. 377). At 5.2ft. an antler pick (No. 380), in a weathered condition; and at lift, some pieces of burnt antler and fragments of decayed oak (No. 397) were obtained. At 4.7ft. from the bottom, an antler pick (No. 399) in good condition was found; it bears clear evidence of cutting in various places, presumably with stone tools. At 1.2ft. above the bottom an antler pick (No. 400), considerably damaged, was un- covered; the beam is very massive, having a maximum circumference of 165mm. between the bez- and trez-tines; circumference just above the burr about 206mm. Within 2ins. of the bottom the crown of an antler (No. 401) of two points was found, somewhat abnormal and flat; it was stained by fire like No. 400.

Above and below the pick (No. 400) a large mass of fragments of charred antler was revealed, some of the pieces being far more calcined than others; one or two fragmentary animal bones were also found here, and a good deal of blackened wood (not true charcoal). This was identified at the Royal Botanic Gardens at Kew as hornbeam (Carpinus betulus). It is said that this tree is now rare in Dorset, though an undoubted native.

Four pieces of flint, much calcined, were found in the mouths of some of the shafts, viz., No. 360 in Shaft XIII., Nos. 893 and 414 in Shaft XIV., and No. 371 in Shaft XV.

In the 1912 Report it was recorded that a remarkable carving in chalk was found in the filling of Shaft X., which perhaps affords further evidence of phallicism in early pre- historic times. This season two somewhat similar objects of chalk were discovered. No. 409 is mentioned in the description of Cutting XXXII. into the earthwork. The other (No. 378) was found in the mouth of Shaft XV., and consists of a piece of carved chalk of circular section, broken at both ends; diam. at larger end 2½ins., tapering to 2⅛ins. at the other end.

Excavation of the Great Bank (Plates II., IV.). — At the same time as the excavation of the enclosure and shafts was in progress, men were employed in penetrating the Great Bank (to the middle of its crest) on the line of the transverse axis. This part of the cutting was 10ft. wide at the top, a considerable batter being necessary to keep the sides standing as the digging was continued downwards.

The upper part of the Civil War terrace, which had to be removed, was composed chiefly of chalk rubble, closely compacted and with horizontal stratification. The maximum thickness of this material, measured vertically, proved to be 5.8ft.; and the width of the whole terrace (at its highest part), measured horizontally, was 33ft. Below it the old surface covered in the XVII. Century was clearly defined, and its junction with the turf at the E. margin of the terrace was quite distinct.

The excavation of the earthwork was carried down in all parts to the surface of the solid chalk; in the contour of the latter there was a considerable amount of irregularity at the W. end, the difference in level being accounted for by a hole, or shelf, in the solid chalk (maximum depth 2ft.), which extended across the cutting from N. to S. (Seen at top of the ladder, Plate IV.) From this hole to the E. end the level of the solid chalk varied to the extent of 1.5ft. As in Cutting XXXII. into the outer part of the Great Bank on the N., the ancient turf line in Cutting XXX., of dark brown unctuous mould (max. thickness 0'5ft.), stood out in marked contrast with the chalk rubble of which the greater part of the earthwork was composed. A length of 25ft. was uncovered, and its surface deviated from a straight line to the extent of 0.75ft. The old surface (represented by a black line in Plate IV.) was discovered at a depth of 15ft. beneath the crest of the bank. Below it the chalk rock was reached at depths varying from 1.65ft. to 2.35ft. Near the E. end, instead of rinding undisturbed rubble under the old surface line, dark mould for a length of 7.3ft. extended down to the chalk rock; no relics were found in this material, and it did not appear to have any special significance, and was probably natural. In one place a dark patch of old turf was met with at a level of 2.2ft. above the old surface line.

The stratification of the chalk rubble and mould forming the body of the bank was extremely interesting, as may be seen on the S. side of the cutting in Plate IV. In the middle there were layers of fine and coarse chalk rubble alternating with narrower seams of mixed mould and fine rubble; these seams were inclined towards the E. at an angle of about 35°. This feature provided an object lesson as to the manner in which the material was thrown up, or carried up in baskets.

At the W. limit of the Great Bank there was a difference in the soils, and the original bank appeared to have been disturbed as far inwards as the old surface line was found to begin. Here (on the S. face of the cutting) the section of what appeared to be a small trench, or ledge, filled with a fine silt of mixed mould and rubble, was observed and plotted; but, as no relics or pottery were found, no definite evidence of date was obtained.

Nothing was obtained from the body of the bank except a few bones of young pig (Sus scroja) near the top, and an antler pick (No. 395), with the handle-end incomplete, found in an important position 1ft. above the old turf line. It is a small shed antler of red-deer with the brow-tine fairly com- plete, and the bez- and trez-tines cut down as stumps. The pick is smooth and bears signs of considerable wear, and is of precisely the same type as those found in the prehistoric shafts.

We obtained no conclusive evidence of the date of the Great Bank; but the fact that few remains were found here and in the similar digging on the N.N.W. (Cutting XXXII.), and that nothing of Roman date was found in either of these cuttings, points to the probability that the earthwork is prehistoric and contemporaneous with the accompanying shafts.

We obtained no conclusive evidence of the date of the Great Bank; but the fact that few remains were found here and in the similar digging on the N.N.W. (Cutting XXXII.), and that nothing of Roman date was found in either of these cuttings, points to the probability that the earthwork is prehistoric and contemporaneous with the accompanying shafts.

(Plate I).

Cutting XXXI. was a quadrilateral area (seen on left-hand side, Plate II.), the sides measuring from 20ft. to 23.75ft. in length. Its N.W. corner abutted against the S.E. margin of Cutting II. Extension. The digging revealed the usual features, viz., the material forming the Civil War terrace resting upon an old surface which covered the Roman and earlier work, the chalk wall of the arena, and the inner and outer trenches with the solid gangway between. On the S.W. the outline of a large shaft (or shafts?), bearing the number XVII., came to light. In the rubble filling at the mouth no prehistoric remains were found, but we examined the material no deeper than 8.15ft. below the surface of the turf.

The E. margin of the pit was practically in the same position as the inner edge of the inner trench. Square post-holes were revealed in two places on the margin of the shaft, one filling the centre of a basin-shaped cavity in the solid chalk. In the inner trench, which was 2ft. deep below the 3ft. gang- way, a circular post-hole was noted. On the arena- wall two strut-holes were found, and two others further S. were trace- able. The wall was uncovered under the terrace to a height of 5ft. above the gangway.

On clearing the outer trench of rubble it was found to have an average depth of 2.2ft. below the gangway. In it the position of six post-holes (mostly rectangular) was clearly traced, the second from the N. being situated in a basin- shaped depression in the bottom of the trench. The post- holes were by no means equi-distant, but they were on an average 3.3ft. apart.

The outer trench was not continuous throughout, and between the second and third post-holes from the N. there was a decided ridge of solid chalk rising from the bottom of the trench. The bottom of the trench terminated at a distance of 3ft. from the S. end of the cutting, and it gradually sloped up to the level of the gangway at 17ft. further S. The discontinuity of this trench was not noticed elsewhere in the excavations.

There was a comparative paucity of relics in this cutting; the unnumbered finds in the Roman deposits included a few iron nails, a stone and an earthenware tessera, and three pieces of terra sigillata. The numbered objects were as follows:-

396. XVI. Century Nuremberg counter, Hans Kravwinckel. Depth 2ft.

403. Third brass Roman coin, burnt and unidentifiable. Depth 4.5ft.

404. Long iron nail, bent, with large thin head. Found on level of the arena-floor.

(Plates I. and V.).

Cutting XXXII. was made through the outer half of the Great Bank on the N.N.W., and was 35ft. in length by 12ft. in width. At the N. end the digging was begun 5ft. beyond the foot of the rampart; from this point the bank rises to a height of 17.4ft. The old turf line was found to be clearly defined for a length of 18.5ft., its max. thickness being 0.25ft. It consisted of a dark brown unctuous mould, without any admixture of chalk. Below its surface the solid chalk was reached at depths varying from 1.8ft. to 2.3ft. The vertical height of the earthwork above the old surface line (clearly seen in Plate V.) was 15ft.

The turf-mould was seldom found to be more than 6ins. thick, and below that most of the thrown-up material con- sisted of chalk rubble, streaked with mould of different colours; some of the lumps of chalk were very large.

In digging the bank at the N. end the outline of the top of a ditch was noticed on both faces of the cutting (Plate V.). On removal of the silting, it was seen that a trench had been cut into the solid chalk subsequently to the formation of the Great Bank, the latter having been considerably scarped in the construction of the ditch. Unfortunately no relics were found in the filling — which appeared to be a comparatively modern accumulation. At the level of the old turf line this ditch was about 7ft. wide, and extended to a depth of 2 - 5ft. below the level of the chalk rock; the ditch was 175ft. wide at the bottom.