Arbor Low Stone Circle [Map] Excavations in 1901 and 1902. By H. St. George Gray (age 29)

The following is an abstract of a paper communicated to the Society of Antiquaries by Mr. Gray, in April, 1903, and printed in Archæologia, Vol. lviii., pp. 461-498. By kind permission of the Society liberal use has been made of Mr. Gray's paper, and the proofs have been revised by him. We are further indebted to the society for the loan of most of the illustrations in Archæologia, but the size of these pages has necessitated considerable reduction of the plan.

These excavations, organized by the Anthropological Section of the British Association, and carried out in 1901 and 1902, were conducted with a view of ascertaining the age of Stone Circles, a beginning, possibly, of a series of such explorations. The actual organization and direction of the work in the field rvas placed in my hands. The ground landlord, the Duke of Rutland (age 83), K.G., the First Commissioner of Works (in whose charge, under the Ancient Monuments Act, the Circle is placed), and the tenant, readily gave their consent for the conducting of the exploration.

In relation to Stone Circles generally, Arbor Low [Map] comes under the heading of those consisting at present of separate megaliths, which, whether single or multiple, are themseives enclosed by an independent vallum and fosse. Other examples are seen at Avebury, in Wilts; at Blisland, in Cornwall, where it has been recorded the fosse is 11 feet (3.35m.) and the vailum 10 feet (3.05m.) wide; at Stennis [Map], in the Orkneys¹, where the vallum is 3 feet (91 cm.) high; and at the neighbouring Ring of Brogar [Map], where the ditch remains, but the wall, which may have been only a stockade, has disappeared². The monument of Arbor Low is very similar in design to "Arthur's Round Table," in Cumberland³, excepting that the vallum of the latter is divided from the fosse by a berme. Arthur's Table, however, does not include any stones.

Note 1. Figured in Fergusson's Rude Stone Monuments, 242.

Note 2. Fergusson's Rude Stone Monuments, 241-2.

Note 3. Op, cit., 128.

Arbor Low is situated in the parish of Bakewell, from which town it is barely five miles distant in a direct line, in a south-westerly direction. The nearest railway station, however, is Parsley Hay, only one mile to the north-west, on the Buxton and Ashbourne branch of its London and North Western Railway. The monument, which is situated on a long ridge of hill, nearly 1,200 feet above the sea-level, commands a most extensive view towards Buxton and Bakewell, in northerly and easterly directions.

Dr. Pegge, writing in 1783¹, says that "the stones formerly stood on end, two and two together, which is very particular." Glover, in his History of the County of Derby (1829), states that "Mr. J. Pilkington was informed that a very old man living in Middleton, remembered when a boy to have seen them standing obliquely upon one end"; tersely adding that "this secondary kind of evidence does not seem entitled to much credit." One of my excavators, an old man, assured me that he had seen five stones standing in his boyhood, and had sheltered under them! On inquiry, however, I ascertained that the man had a reputation for gross exaggeration. The Rev. S. Isaacson, writing in 1845, was of opinion that "these stones were never placed in an erect position." He says further that "the imported stones all appear to be resting on the native rock, the comparatively thin covering of soil having accumulated through the lapse of centuries." Gardner Wilkinson, on the other hand, in 1860, says "it is evident that they originally stood upright, as in other sacred circles." Lord Avebury, writing some twenty-five years ago, stated cautiously, "It is doubtful whether they were ever upright."² As recently as 1899, Dr. Brushfield appeared to be of opinion that the stones originally stood upright; and Mr. A. L. Lewis is of the same opinion³.

Note 1. Archeologia, vii., r4z

Note 2. The Reliquary, xx. 81-85, with view and woodcuts.

Note 3. Man. (Anthropological Institute), September, 1903, No. 76.

The published plans of Arbor Low are for the most part far from correct, Sir Gardner Wilkinson's plan, figured in two places, being the only exception¹. In this small plan the position of the circle of stones is fairly correct, although there are several discrepancies in the proportional size of the stones. Pegge represents the megaliths as forming a true circle, and speaks of his plan as being" sufficiently accurate²." Glover includes what is styled "an engraving copied from an accurate³ drawing by Mr. Mitchell⁴." It could not well be more inaccurate.

Note 1. Journal of the British Archaological Association, xvi., plate 9; and Fergusson's Rude Stone Monunenls, 140.

Note 2. Archaologia, vii., r48.

Note 3. The italics are mine.

Note 4. Stephen Glovet, History of the County of Derby, i. 275-6.

The area or plateau enclosed by the fosse presents an uneven surface, but the contours across this part of the plan have been delineated to follow the general slope of the ground, and not to mark, in this part, every little depression or slight elevation as it occurred. The contours of 6 inches vertical height show the shape of the monument and its immediate surroundings within the "square¹." The highest contour comes on the top of the tumulus on the south-east rampart; the lowest, at the north corner of the survey, a fall of 24½ feet (7.47 m.) in the ground from top to lowest part.

Note 1. To ensure absolute precision, eighteen hundred levels were taken.

The average height of the vallum above the general surrounding turf-level, as shown by the contours, is, on the outside, 7 feet (2.13 m.); its height above the central plateau, about 6 feet (1.83 m.). The average vertical height from the crest of the vallum to the surface of the silting of the fosse is 12 feet (3.65 m.). These dimensions have been much exaggerated by successive writers on Arbor Low¹; and as recently as 1900 the average height of the vallum has been recorded as 16 feet (4.88 m.) above the surrounding level².

Note 1. Archæologia, vii., 142.

Note 2. Joarnal, British Archæological Association, N.S..vi., 129.

Judging from those portions of the fosse already re-excavated, the material obtained from the fosse when it was first excavated was not enough to complete the construction of the vallum. The confines of the vallum are bounded at various points by ten small govemmental stones, indicated by small black squares on the plan. The continuation of the ditch and rampart is interrupted on the north-west and south-east by the entrance causeways, which are not in line with the central group of stones. The causeways are on the same general level as the area occupied by the megaliths and the surrounding land. The circumference of the rampart, including the entrances, is about 808 feet (246 m.).

The vallum is joined on the south-west by a slightly raised bank about 1½ foot.(45.7 cm.) high, and an almost imperceptible "silted-up" ditch, which run for some distance in a southerly direction, and about which there have been various theories. Some writers have connected this so-called "serpent", with Gib Hill [Map], a tumulus at a distance of 1,043 feet (318 m.¹) from the centre of Arbor Low (plate I.). Gib Hill [Map] was unsuccessfully dug into about 1812, and again by William Bateman in 1824², when a few stone implements appear to have been found. In January, 1848, Thomas Bateman made a more thorough examination of the mound, when he discovered a cist, the top stone of which was only 18 inches (45.7 m.) from the apex of the tumulus, containing a cremated interment in a small urn of Bronze Age type³.

Note 1. According to my tape measurement.

Note 2. Bateman's Vestiges of the Antiquities of Derbyshire, 31. [Map]

Note 3. Bateman's Ten Years' Digging 17-20

On the south-east, adjoining the external face of the vallum and partly resting on it, a tumulus [Map] stands, the summit 7½ feet above the surrounding turf-level (see photograph, plate II.). This barrow was first attacked in 1770 by the then occupier of the farm, without success. Likewise to 1782 by Major Rooke, assisted by John Manders, and in 1824 by William Bateman and Samuel Mitchell, of Sheffield. A fourth attempt, made in 1845, by Thomas Bateman and Rev. S. Isaacson, resulted in the discovery of a limestone cist, which has been frequently described¹. It contained calcined human bones, a bone pin², pyrites and flint, and two small urns³, differing considerably in style and ornamentation, but undoubtedly of Bronze Age manufacture, and probably rather early in that period⁴.

Note 1. "Arbor Low," by Sir John Lubbock, the Reliquary, xx. 8l-85; Bateman's Vestiges of the Antiquities of Derbyshire, 64-66 and 74; and Winchester Volume of the British Archrological Association (1845), 197-204.

Note 2. Figured in Fergusson's Rude Stone Monuments, r4r, and Vestiges, 65,

Note 3. These relics are in the Sheffield Museum. The urns are reproduced by kind permission of Mr. E. Howarth, the curator.

Note 4. Dr. Brushfield calls my attention to the very misleading representation of this urn in Vestiges of the Antiquities of Derbyshire, p. 65.-Ed. D.A.N.E S

These urns are figured in the accompanying illustrations, figs. 1 and 2. Fig. 1 was found filled with burnt bones. It is unusually wide and low: 4½ inches high, 9 inches diameter at top, 4 inches diameter at base. The other pot, fig. 2, found with it, is 4½ inches high, 5½ inches diameter at top, 3 inches at bottom. In June, 1845, the digging of this tumulus [Map] was resumed, but nothing further was discovered beyond a few pieces of deer-horn. Mr. Bateman never took the trouble to fill-in his excavation properly, the result being that four or five little knolls exist round the top of the tumulus, bounding a rather deep depression in the centre (see photograph, plate II.). In addition to this, he threw some of his débris into the ditch, clearly shown in the plan and photograph plate I. The formation of this tumulus, which is probably of somewhat later date than the vallum, has caused a gap to occur in the vallum on either side of the mound. There is also another irregularity in the form of the rampart to the north of the tumulus, caused by a kind of spur which extends halfway across the fosse. All along the crest of the eastern and north-eastern vallum are irregular depressions, sufficient material for filling which may be observed at intervals in ledges and patches along the base of the inner side of the east and north-east vallum, or, in other words, along the outer edge of the fosse in these parts. The only feasible explanation for this seems to be that Messrs. Bateman and Isaacson, elated by their success in finding the interment in the tumulus close to, pursued their investigations along the adjacent crest of the vallum at intervals, shovelling the material inwards down the slope of the rampart.

Section 2, 10 feet (3.05m.) wide, was next commenced on the north-western side at a distance of 16 feet (4.88 m.) to thewest of the north-by-west causewa)'. (See section line E-F on plan.) The greatest depth of the ditch in this cutting was found to be only 2½, feet (76.2 cm.); and nothing being found here, the re-excavation of the ditch was continued¹ from this point in the tlirection of the causeway, the solid sides of which were found. Stone relies were fairly numerous in this part, called "Ditch Extension, Section 2." The average depth of the ditch, the bottom of which was very uneven, was 3 feet (91.4 cm.) here.

See photograph, plate viii.

Amongst the relics found here were six flakes¹ of black flint, of fine quality, weathered white to a consicierable depth, mostly of exceptionally large size; they lay together, at a depth of 2.7 feet (82 cm.) from the surface, on a ledge on the solid side of the causeway (No. 17, plan, and section line N-K). These flakes must have been placed by hand on the ledge and forgotten, eventually becoming buried in the silting, or perhaps purposely concealed; they could not have come by accident into the position in which they were found. The flakes are of considerable size and weight. Though of irregular form, their edges are still sharp and undamaged. Mr. Henry Balfour has suggested that they might possibly have been intended to be used as digging tools. They are figured in fig. 7, page 65, ⅜ scale linear.

Two of these flakes have since been found to join.

There were also found in the Ditch Extension of Section 2, the flint implements numbered 5, 8, 10, 11, 13, 14 and 15 on plate V., and similarly marked on the plan.

In this cutting a small oval-shaped hole in the limestone floor was found, filled with a stiff clayey mould, but no relics were discovered in it. The only animal remain here was a tooth of sheep, depth 0.7 foot (21 cm.)

Section 3, also 10 feet (3.05m.) wide, was a cutting across the ditch, midway between Sections 1 and 2, on the west side. The silting was very soon removed in this case, the uneven limestone floor being found at a maximum depth of 1.9 foot (57 cm.). Three stone implements were found here, viz., Nos. 3, 6, and 7 on plate V. and plan.

Section 4 and its extension to the west to find the limestone sides and enrl of the ditch at the north-north-west causeway proved to be the most interesting and productive of the six ditch cuttings. (See photograph, plate VII.; and section lines J-L, Q-R, and Y-2, on plan.) The first "find" was an almost circular greyish-white chert end-scraper ("28" on plan, and plate VI., No. 28). One of the side edges exhibits some fine secondary chipping, and woulcl serve admirably as a knife. At a depth of 3 feet (9r cm.) closc to the limestone side of the ditch, at the south-east of the section, about two-thirds of an extremely thin and finely chipped flint arrow-hearl were found; its greatest width is 21 mm., greatest thickness 2.7 mm. Its position is shown in a photograph of the cutting (plate III). The base of this delicate implement is bounded by a semi-circular arch, whilst the side-edges in the perfect arrow-head, from the points of greatest width to the tip, appear to have been quite straight, thus representing the lozenge-shaped form as regards the upper half and the leaf-shaped variety in the lower half ("29" on plan, and plate VI., No. z9). This type belongs to the "kite-shaped class" of Mr. W. J. Knowles.¹

Note. Although belonging to this class, Mr. Knowles does not figure an arrow-head precisely similar. Journal of the Anthrolological Institute, xxxiii. 44-56.

On the bottom of the fosse, at a depth of 5.7 feet (1.74 m.), under stiff clayey mould, was found the most interesting relic that Arbor Low has yielded during these excavations, viz. ("43" on plan, and plate VI., No. 43), a barbed and tanged chipped arrow-head, of greyish-white flint or chert, of extremely symmetrical form and 1 inch (25 mm.) long, a small portion only of the tip being deficient; greatest width, 21.5 mm.; greatest thickness, 5 mm. it is finely chipped all over its surface, including tangs and stem, which latter are bevelled on both faces from median ridges to edges, The end of the tang, which is very slightly longer than the barbs, is also bevelled. The section across the arrow-head is bi-convex, but it is considerably flatter on one face than on the other. The cutting edges are slightly convex, owing to the fact that the barbs have an inward curve.

Not far from this beautiful arrow-head a small rough chert scraper was found (plan, and plate VI., No. 37). In clayey-mould, close to, at a depth of feet 4.6 feet (1.40 m.), the greater part of a small horn of red-deer, with four tines, in an extremely friable condition, was discovered; indeed, only a portion of one tine could be preserved ("39" on plan).

In the extension cutting to the west a greyish-brown flint flake, with secondary chipping, was found at a depth of 12 foot (36.5 cm.), "40" on plan, and plate VI. At "41" (plan), a small chert flake was found, depth 12 cm. Of animal remains, a tooth of sheep was found at a depth of 3 feet (91 cm.), portion of a humerus of ox, at a depth of 5 feet (1.52 m.), and at various depths several small fragments of animal bones, too minute and friable for identification.

Section 5, 10 feet (3•05 m.) wide, is situated on the west between Sections 2 and 3 (plan). Like Section 3, it proved to be very shallow, the limestone rock being reached in the middle of the silting at a depth of 1 foot (30 cm.) from the surface, whilst on the inner side of the ditch the depth was only foot (24 cm.). The bottom was fairly level, especially when compared with the bottom of the fosse in all other parts except in Section 3. No relics were found in Section 5. The photograph, plate II., taken from the north-west vallum, shows, in the foreground in the right-hand corner, the position of this section in its re-turfed condition at the close of the excavations.

On the east by north, another cutting, 10 feet (3.05 m.) wide, Section 6, was made across the fosse. In this part, the surface of the turf of the silting at the lowest part was 7 feet (2.13 m.) lower than the general level of the ground immediately adjacent to the central stones, as shown by the contours in the plan. Here again nothing was discovered, with the exception of a few very small unidentifiable fragments of animal remains too small for identification and therefore not preserved. The average depth of the cutting below the surface of the silting was 4.3 feet (1.31 m.), and the maximum depth 5.8 feet (1.79 m.) in the north-west of the section. A block of limestone rises to a height of only 7 inches (18 cm.) below the surface of the silting. On the east and north-east the median block of limestone which ran across the section was divided from the limestone side of the ditch on the east by a vein of fine clay, yellowish-brown on the top and white below, which occurred at an average depth of 4 feeet (1.22 m.) from the surface of the silting; the surface mould reached a depth of 2 feet (61 cm.), below which was the usual stiff clayey mould thickly mixed with small pieces of chert; at the bottom the proportion of clay to mould itlcreased, and the silting became more stiff and rnoist; in fact at the bottom it had to be cut out in solid lumps with a small spade, the chances of finding relics consequently being very remote.

Considerable traces of fire were observed at the bottom of this cutting, especially in the south-east corner, indicated by a dark patch near the bottom of the photograph, plate III. Dr. Garson has suggested that the traces of fire at this point may possibly indicate that this portion of the ditch, which is deep here, and therefore sheltered from winds, was occupied by persons employed in guarding the circle; hence the greater number of implements in this section and the corresponding section, No, 2, on the other side of the north-north-west causeway.

The completion of the excavation of the Ditch Extensions of Sections 2 and 4 enabled me to determine the minimum width of the northern causeway as 29 feet (8.84 m.). This causeway is partly shown on the left-hand side of the photograph, plate VII. The southern causeway is well shown in plate I.

This completed the examination of the fosse, 85 feet (26 m.) having been excavated in all, of the total length of 540 feet (165 m.). In other words, nearly one-sixth of the fosse has been re-excavated, flint implements only having been found from top to bottom of the silting. The mean depth of the whole fosse excavated is 3.9 feet (1.18 m.), and the average width at top 22 feet (6.7 m.).

As regards the arrow-heads, it is worthy of notice that the barbed and tanged specimen (plate VI., No. 43), a form generally considered to be the most highiy developed, was found 2.7 feet (82 cm.) lower down in the silting than the arrow-head of leaf-shaped form, but approximating closely to the lozenge-shaped, a form rvhich is generally regarded as an earlier form than the barbed variety (plate VI., No. 29). It has not, however, yet been clearly ascertaiued which form of arrow-head was first manufactured, and the matter is at present surrounded with some difficulty, inasmuch as the triangular, stemmned, and leaf-shaped varieties have been placed ou record as being found together in the same locality and the same deposits. One form is easily evolved from another, and although General Pitts-Rivers' method¹ of arranging arrow-heads, showing the theoretical transition from one form to another, is excellent in museum arrangement until something more definite is arrived at, yet, bearing in mind the records of the circumstances of the finding of flint arrow-heads during the last thirty-five years, it would be, as Sir John Evans has said long age, "unwarrantable to attempt any chronological arrangement founded upon mere form, as there is little doubt of the whole of these varieties having been in use in one and the same district at the same time, the shape being to some extent adapted to the flake of flint from which the arrow-heads were made, and to some extent to the Purposes which the arrows were to serve."²

Note 1. Colonel A. Lane Fox's second lecture on "Primitive Warfare," Journal of the Royal United Service Institution, 1868, xii., No. li.

Note 2. Ancient Stone Implements of Great Britain (1872), 330; second edition, 370.

The two Arbor Low specimens were probably not used a great many years apart, for the fosse would, throughout its lower portions, and indeed within a foot or two of the present surface of the silting, fill up somewhat rapidly, particularly at the bottom, owing to the fact of the sides of the ditch being exposed to the erosive action of the weather, and the consequent disintegration of the sides.

The barbed arrow-head being on the bottom of the ditch and near the middle would become covered almost irnmediately the fosse was allowed to silt up. The other arrow head being found within an inch or two of the side of the ditch at a depth of 3 feet, it is obvious that it would be deposited on the talus and become covered very soon after the barbed arrow-head. As before mentioned, the broken kite-shaped arrow-head was picked up out of the silting so very near the actual wall of tl're ftrsse that it is just possible it may have rested on a small ledge of the lirnestorle rock, being removed therefrom by the pickaxe on the day of discovery. In any case it may, I think, be safely asserted that these arrow-heads were in use at about the same period, suggesting that these forms were probably contemporaneous in this district. Under these circumstances, this kite-shaped arrow-head mighb be regarded as a development of the typical leaf and lozenge-shaped forms, made about the same time as the barbed and tanged specimen, which latter, as a type, is considered to be a late development in the art of the manufacture of the stone arrow-head; although stems are known to have been developed in the Palaolithic Age.¹

Note 1. Journal of the Anthropological lnstitute, xxxiii. 52, 54.

In continuation of Section 2 across the fosse a cutting, 10 feet (3.05 m.) wide, was made through the vallum. This point was chosen as the vallum was rather low here and consequently would not entail so much labour. No relics were found in this cutting, except a doubtfully-artificial stone scraper picked up on the old surface line ("4" on plan, and plate V.)¹. The absence of relics in this section was very disappointing. The cutting, however, was of value in showing the material out of which the vallum was constructed. (See photograph of the cutting, plate VIII.) Measururing from the crest of the rampart downwards, the soils, etc., occurred as follows: (1) turf and turf-mould, 0.5 foot (15 cm.); (2) rough pieces of thin-bedded limestone mixed with a little mould, 3.3 feet (98 cm.); (3) band of small pieces of chert with a little mould, 0.3 foot (9 cm.); (4) yellowish-brown clayey mould, 0.5 foot (15 cm.); (5) "old surface line" of dark brown mould, 0.3 foot (9 cm.); (6) light yellowish-brown sand.

Note 1. The surface of chert does not seem to alter as flint does from exposure and age; consequently it is often difficult, if not impossible, to decide whether certain fractures are ancient or qurite recent.

Like Section 2 of 1901, the section through the vallum in continuation of Section 4 yielded no relics, but proved of interest, inasmuch as it was largely composed of huge boulders of limestone, maximum length 4 feet (1.22 m.), as shown in boulders had been loosened in the formation of the fosse and utilised for the construction of the vallum. The crest of the rampart here is about 4'7 feet (1.43 m.) above the "old surface line" immediately below it. Chert and calcite occurred in bands in the limestone strata in large quantities here, and fluor-spar was detected by Professor Boyd Dawkins.

These excavations were made with a view of ascertaining whether holes existed in the limestone floor in which Stones I., IL, III., and XXXVII. originally stood, but the results could hardly be considered conclusive.¹ As stated before, the various writers on Arbor Low disagree as to whether the stones originally stood in an upright position, or whether they always lay flat on the ground.

Note 1. More time could not be bestowed on this particular investigation.

A hoie, 7 feet by 4½ feet (2.13 m. by 1.37 m.), was made to the east of, and close to, Stone XXXVII. Therc was a well-marked depression in the turf here, and the stone has a flat squared surface at this end. Just below the turf, depth 0.3 foot (9 cm.), a small fragment of Romnano-British pottery was found (pian "35"), and a small discoidal flint scraper in mould at a depth of 0.6 foot (18 cm.) ("34" on plan, and plate VI., No. 34. A hole in the limestome floor certainly existed close to the north-east of the stoue, of more or less oblong form, length 5'8 feet (1.79 m.), maximum depth below surface 2.1 feet (64 cm.)1 it, however, appeared to me to be merely one of the usual natural shallow depressions in the limestone, and the excavation afforded no evidence of a hole having been cut for the reception of the base of a standing stone. Another hole 8 feet by 4½ feet (2.44 m. by 1.37 m.) was made at the north-west end of Stone II. No evidence as regards the original position of this stone, if it ever stood upright, was adduced from this digging, and nothing was discovered but a flint flake (plan "38"), depth 0.5 foot (15 cm.).

The south-west edge of Stone I. in the centre being bordered by a marked trench, an excavation 13 feet by 5 feet (4 m by 1.52 m.) was made. Digging had evidently taken place here in recent times, probably by Messrs. Bateman and Isaacson in 1845, when they explored the tumulus on the south-east vallum. The rock was reached at a maximum depth of 61 cm. below the highest part of the turf, and a minimum depth of 30 cm. Early Victorian shards were found, and a clay tobacco pipe bowl, nineteenth century; also a flint flake, depth 9 cm. (plan "36"), and a fragment of pottery, apparently of Romano-British quality ("42" on plan), depth 40 cm.

A larger piece of trenching, 35 feet by 7 feet (10.67 by 2.13 m.) was dug in 1901, between Stones I. and II., and to the east of them in the direction of Stone XXXVII. To the west a stump (No. 13) was found just under the turf, standing in a leaning position towards the north-east.¹ Between these stones, at a depth of 0.5 foot (15 cm.), another fragment of Romano-British pottery² was found ("19" on plan), and at a depth of 0.9 foot (27 cm.) a small chipped flint implement ("20" on plan), length 33 mm., width 28 mm., resembling a broad leaf-shaped arrow-head in form, flaked on both sides; but it may have been hafted and used perhaps as a knife, as the point is extremely obtuse, and therefore not well adapted for penetration. It is figured in plate V., No. 20.

Note 1. This was left in situ at the completion of the excavations,

Note 2. It consisted of a fragment of rim, grey on the outside and brick-red on the inslde.

No holes for the reception of the bases of the monoliths were found between these stones. In fact, the undisturbed ground in this part was struck at about 1.7 foot (52 cm.) from the surface. The present evidence afforded by the digging of these holes seems hardly sufficient to warrant the assertion that the builders of the stone circle did not cut holes in the limestone for the reception of the stones, that is, supposing they originally stood upright.

The approximate length of the left femur is about 453 mm., which gives a stature (by Rollet's method) of 5 feet 5 inches (1.65 m.). The skull has been restored as far as possible, and turns out to be mesaticephalic, or medium-headed, with a cephalic index of about 780. At the Glasgow meeting of the British Association, Dr. Garson read, a detailed report on the osteological this skull, concluding by saying that "the osteological characters show that the individual was not of the type found in interments of the Neolithic period, neither do they point to his being of the Bronze Age type, though he was more nearly allied to it than to the former. On the other hand, there are no characters about the specimen which would preclude its being much more recent. The extended position in which the body had been laid decidedly supports the view of the interment being of more recent date than the Bronze period, to which I consider the weight of the evidence afforded by the osteological characters also points."

Note 1. The centre of the circle is marl

Excavations In The Small Dyke Connected With The Vallum Of Arbor Low On The South-West. (See Plan, Page 41.)

The primary idea of excavating here was to ascertain whether the ditch of this small dyke continued under the rampart of Arbor Low itself, and if possible to prove its age by means of any relics that might be found. The ditch on the surface is marked only by a very slight depression, and the bank at the present day averages only foot (457 cm.) above the level of the surrounding field. Doubtless the bank was higher at the time of construction, but denudation has intervened, and loss by material gradually sliding down, thus assisting in the formation of the silting of the ditch. A section, 7 feet (2.13 m.) wide, was first made through the bank and ditch at a point 170 feet (51.82 m.) from the centre of Arbor Low, and within the area of the plan. The finds in the ditch here were: a small worked flake of yellowish-brown colour and translucent, depth foot (37 cm.), "22" on plan, and the accompanying section, fig. 8; and a flint flake, "24" on plan and section. The depth foot (55 cm.), bank yielded: At "27" on plan and section, depth 0.9 foot (27 cm.), on the level of the "old surface line," a well-formed greyish-white chert (or flint) end-scraper, exhibiting signs of considerable use. It is figured in plate VI., No. 27. At "30" on plan, and section, depth 0.7 foot (21 cm.), a greyish flint scraper, very smooth and worn, indicating prolonged use. It is figured in plate VI., No. 30. These implements (Nos. 27 and 30) are undoubtedly of the date of the construction of the dyke.

No further excavation of the bank was made, but the exploration of the ditch was extended for 4.5 feet (1.37 m.) to the north, and produced the following stone objects: At "25" on plan and section, a long greyish-white chert flake, with rough and irregular serrations along both edges, depth 2 feet (61 cm.); the oblique top does not appear to have been worked; the bulb of percussion displays a large and well-marked êraillure. At "26" on plan and section, at a depth of feet (98 cm.) close to the bottom of the ditch, a small white flint knife, finely chipped and of somewhat triangular form, length 30 mm., width 20.5 mm., greatest thickness 7 mm., was discovered. It is of an uncommon form, and at first sight would probably be included in the category of arrorv-heads; it is, however, apparently a finished implement, being chipped all round the edges; the edge at the base shows signs of crushing or bruising; one of the side-edges is straight and neatly chipped, whereas the other is convex with a finely-rvorked bevelled cutting-edge with signs of crushing near the base. On the other face the concave edge has been considerably worked; the bulb of percussion has also been chipped, leaving part of the êraillure facet visible. This is the most important relic found outside the vallum of Arbor Low during these explorations; it is figured in plate VI., No.26. Two flakes were also found here both at a depth of 2 feet (61 cm.): one is merely a long, narrow, outside flake; the other a flint flake - edges slightly serrated.

At this point the little bank had been levelled down, probably for a modern track, re-commencing and almost immediately terminating in the vallum of Arbor Low. Here another small excavation, 8 feet (2,44 m.) wide, was made into the ditch to detemine whether it ended here or continued in a northerly direction under the vallum of Arbor Low. The itch, however, shelved up gradually, and the rock-end was found (as shown in the annexed section, fig. 10, on line V-X of plan), thus proving that this little earthwork is of the same period of construction as Arbor Low or later; but judging from the relics discovered, it would appear to be of about the same date as the larger monument. In this latter excavation a calcined chert scraper was found, 0.9 foot (27 cm.) deep ("31" on plan, section, and plate VI.). At "32" on plan and section (No. 32, plate VI.), a small, elongated, narrow black chert end-scraper, worked on both sides, was found, depth 1.6 foot (49 cm.); and at "33" a black chert flake,depth 1.8 foot (55 cm.).

The average depth of this little ditch beneath the surface was 3 feet (91 cm.); the width at top, 8 feet (2.44 m.)

During the four weeks that the excavations were in progress, no metals were discovered, nor any traces of fictile ware that could be assigned to the date of the construction of Arbor Low. Six sections were cut through the fosse, 85 feet (26 m.) of fosse in all; two cuttings were made through the vallum; four patches of trenching of varying dimensions were dug in the interior; and of the small dyke to the south-south-west, 7 feet (2.13 m.) of rampart and 19½ feet (5.94 m') of ditch were examined. The number of relics found¹ has certainly been disappointing, and yet, on the other hand, Arbor Low, not having been a habitation, and from negative evidence appearing not to have been a place of sepulture at a period closely following its construction, I do not know that more relics could be expected under the circumstances. Nothing Roman has been found, except three small fragments of what appears to be Romano-British pottery just below the turf in the interior.

Note 1. The relics have been presented by the British Association to the British Museum, and the writer's Model of Arbor Low, fully described and figured in "Man," October, 1903, No. 84, p. 145, will also probably find a home there.

As is well known, flint scrapers are frequently found during successive ages down to and including Roman times, but here they are found deep in the silting of the fosse, only in association with other rude stone implements and chipped flint arrow-heads of Neolithic form, but of a variety found also in later periods. The majority of the implements fouud at Arbor Low appear to be of chert, which is only what would be expected, seeing that it is indigenous and an excellent substitute for flint, whereas flint was brought probably from some considerable distance.

Arbor Low is of such precise age as the barbed arrow-head may be assigned to, it having been found on the bottom of the deepest portion of the fosse. This form, being usually recognized as a late development in Neolithic flint working, points to the probability of the construction of the fosse and vallum not being assigned to a date earlier than the Late Neolithic period, although, judging from various "finds" of Neolithic arrow-heads of barbed form, they may perhaps have been in use, in some districts, about the middle of the Neolithic period. However, there were no indications that this arrow-head might have reached its position at a date long subsequent to the formation of the fosse. On the other hand, it is well known that the "tanged and barbed" type of arrow-head is very frequently found associated with Bronze Age finds. The existence of stone implements to the exclusion of bronze does not necessarily establish a Neolithic age for a monument such as this; and yet, considering the amount of excavation done, the absence of bronze has some significance. If Arbor Low was actually constructed in the Early Bronze Age, we should not expect to find any abundance of the then precious metal. Given small fragments of broken bronze would not be thrown away at this period; indeed, every care would be exercised to preserve fragments for the melting-pot, much as we preserve all broken and disused gold objects, merely on account of their value as gold.

On the south-east vallum a, Bronze Age tumulus [Map] was constructed undoubtedly from material derived from the original monument of Arbor Low. As previously stated, no bronze was found here or in Gib Hill [Map], just over 1,000 feet distant; but their other contents point to a Bronze Age culture, probably not particularly late. If the "finds" from this tumulus¹ on the vallum of Arbor Low are to be regarded as belonging to the Early Bronze period, "then," as Mr. Henry Balfour said at Belfast, "the probability of the circle being of Neolithic date is much increased."

Note 1. The absence of bronze in this interment does not necessarily give an Early Bronze Age date for the burial, for bronze was rarely found with interments even in the fully-developed Bronze Age. The two pots found in the tumulus (figs. 1 and 2) are not "beakers," or drinking-vessels, which the Hon. John Abercromby has recently classified as being the oldest Bronze Age ceramic type in Britain, Journal of the Anthrolologieal Institute, xxxii. 373.)

The absence of finds on the old surface line under the vallum in the parts examined unfortunately does not help towards the solution of the problem of date. Our conclusions, in the present state of our knowledge, have to be deduced on somewhat meagre evidence as regards the quantity and nature of the relics found. The discoveries made in 1902 correspond in the main with those made in 1901.

In conclusion, it remains to be said that the date of the construction of the Arbor Low Stone Circle should be located, in accordance with the evidence derived from these explorations, within the period covered by the Late Neolithic and Early Bronze periods; in other words, the period of transition from stone to bronze.¹

Note 1. The date of construction of Arbor Low appears to tally precisely with Mr. Gowland's deductions as to the date of the erection of Stonehenge, from evidence derived frorn his excavations there in 1901. (Archaologia, lviii. 85, 86.)