THE ST. NICHOLAS CHAMBERED TUMULUS, GLAMORGAN By JOHN WARD, F.S.A., Curator of the Cardiff Collections, National Museum of Wales

[The St. Nicholas Chambered Tumulus is now referred to as Tinkinswood Burial Chamber [Map]]

St. Nicholas is a small and ancient village, six miles west-south-west of Cardiff, on the highway from that city to Cowbridge, Neath and Carmarthen-a well-known road of Roman ancestry, and popularly known as the Via Julia. A mile-and-a-quarter to the south of the village is Dyffryn House¹, the residence of Miss Cory and midway between the two, in a small plantation near Tinkinswood farmhouse, are the remains which are the subject of this paper. Plantation and farmhouse are on the apex of a wedge-shaped spur from the high ground at St. Nicholas. Along the south-west foot of this spur flows the Weycock, locally known as The Brook, a considerable stream which below Dyffryn House veers to the west and eventually flows into the estuary of the Dawe at Aberthaw while along the eastern foot is a small tributary which joins the parent stream a quarter-of-a-mile below the farmhouse. The summit of the apex is flat and about 280 ft. above the sea-level, with the plantation on its south-western brow and the farmhouse near its eastern. On its northern side, this small plateau is defined by a fall of several feet, beyond which the ground gradually ascends to 380 ft. at St. Nicholas. The surroundings are pleasantly diversified with hill and dale, woodland and pasture.

Note 1. Now usually spelled 'Duffryn.'

The valley of the Weycock, from its head near Cottrel, half-a-mile west of St. Nicholas, to about a mile below Dyffryn House, is known as Dyffryn Golych, Golych being the ancient name of the stream, or at least this portion of it. It is mentioned under this name in Liber Landavensis, as also a place called Tref-Gulych, which is generally identified with St. Lythans, but is more likely to have been nearer to, if not close by, the stream. Possibly it was the Welsh name for Worlton, which lay somewhere a little to the south of Dyffryn House, and the few houses now known as Dyffryn village may represent it. Early in the fifteenth century, Worlton passed by marriage to the Buttons, of which family Admiral Sir Thomas Button was a distinguished member and as there is a homestead moat close by the village, it is reasonable to think that it was the site of their house. Dyffryn House appears to have become the residence about the beginning of the eighteenth century.

The map, Fig. 1, presents the site of the chambered mound (A) and its immediate surroundings. It will be noticed that both the plantation and the adjoining field on the east, are known as Caste! Carreg (' stone castle'), and this is suggestive that the two were formerly a single enclosure. Such a name could not at the start have been given to a piece of ground but, as will be seen later, it was a popular designation of the megalithic structure upon it. This appears now to be forgotten, for whenever I have asked the local people what they call this structure, the invariable reply has been, The Cromlech.' Crossing the field is the foot-path between St. Nicholas and Dyffryn. Further to the east and at the foot of the field known as the Stable, is the small tributary of the Weycock mentioned above. The only other field to be noted is that called Grindstone, on the south. The dot-and-dash line indicates the limits of the plateau.

The capstone has long been regarded as the largest in Great Britain, and there is no reason to doubt this. It is, however, surpassed by certainly one Irish example, that of the great Kernaston chamber in Co. Carlow, the dimensions of which are stated to be 23 ft. 6 in. by 22 ft., and weight 100 tons. On account of its relative thinness, our capstone is probably not the heaviest in Great Britain. Its weight has been estimated as about 50 tons but my own estimate, based upon the cubic measurements and specific gravity, is considerably less. The specific gravities as determined by Mr. North, of three chippings from the slab, were 2.49, 2.64 and 2.67, and I took 2.60 as the average. I made two calculations of the bulk of the slab, the first being 546 cubic feet, and the second 552 cubic feet. From these two calculations I obtained the results, 39.6 tons and 39.93 tons. The latter, I have reason to think, is the nearer approximation to the truth and as Mr. North considers that the chippings, necessarily taken from the surface of the stone, would, through the action of weathering, be less dense than the average, it is almost certain that this is an under-estimate of the weight.

Literary References and Folk-Lore. The earliest literary reference known to me, is in a paper, Observations on a Tour through South Wales," etc., communicated by Owen Salusbury Brereton to the Society of Antiquaries in 1772 (" Arehseologia," III, p. 116) About eight miles eastward is the parish of St. Nicholas, where Mr. Price of Duffrin lives, in whose grounds are three fine Kistvaens or Cromlechs they all consist of four stones of immense size the cover stone of the first is full 6 yards long, and about 5 broad and 21 in. thick the back stone is about 5 yards broad it is called Carrig maen Llwyth. The second is in the next field, rather less in dimension and sunk low in ground, but exactly of the same form. The third is a few fields more eastward, quite perfect, in its ancient form, and full 8 ft. high it is called Guael-y-Velin, and sometimes Maes-y-Velyn, from the field it stands in. All three stand on high ground, open to the west, and are in full view of each other. A coverstone of a fourth stands upright in a field close to the roadside between Bolston and St. Nicholas." Of these remains, the first and third are our two chambers the second is one of the two piles of stones, c and D on our map and the fourth, the large slab in a field on the south side of the main road nearly opposite the gates of Cotterell Park, which is marked Standing Stone on the Ordnance Survey. Brereton evidently did not understand Welsh. Probably what is intended by Carrig-maen-Llwyth is 'grey stone'; and there is little doubt that his Gwael-y-Velin' is a blunder for Gwâl-y-filast, 'the kennel of the greyhound bitch.' His statements are repeated, only in condensed form, in King's Monumenta Antiqua," 1799 (Vol. I, p. 242).

Dr. Malkin's description of the remains in his Scenery, Antiquities, and Biography of South Wales," 1807 (Vol. I, pp. 164-175) is longer and more accurate. He described the St. Nicholas chamber as a rude piece of antiquity," consisting of five upright stones and a horizontal one, 24 ft. in length and 17 ft. "in its widest part "-measurements which exceed the actual dimensions nearly as much as Brereton's came short. He noted the vestiges of the mound, which he estimated to occupy something like four square perches of ground," and conjectured to have originally covered the chamber. His statements show that the remains were in much the same condition as imme diately before the recent excavations.

About fifty yards south, in another field, appear the quarry and rock whence these large stones were taken. There are in the same field with the quarry two flat stones of no inconsiderable, though comparatively small, dimensions, set up nearly in the same manner, though much encumbered by rubbish and brambles and the stile, by which the two fields communicate, is formed by a very massy piece of the same stone, which seems to have served that purpose for ages." It is not easy to follow Malkin here. Fifty yards south of the chamber is the upper end of the field called Grindstone, but there is no sign of a quarry there, unless it is a slight hollow which very conceivably was caused by the removal of stone at some remote period; it is, however, not large enough to have yielded the enormous slabs of the chamber. Probably he really meant east, and referred to what is now known as The Quarry. Mr. North has compared the slabs with the exposed rock of this spot, and is strongly of opinion that it is their source. There are large gaps in the stratum, such as would be left by the removal of the slabs, and the transportation of these would be comparatively easy, as the intervening ground is level. Malkin's two flat stones set on end exactly agree with the two marked F on our map, except that these are not in the same field with the quarry." They may, however, have been at the time, as the field-fences hereabouts have since been altered. His stile clearly is the stone E, close by the present wooden stile.

Malkin also described with considerable care the St. Lythans chamber-the 'greyhound-bitch kennel,' as he rendered the old Welsh name for it. "It is rather singular," he remarked, that these cromlechs .... should almost everywhere in Glamorganshire .... be known by this uncouth term of greyhound-bitch kennels." He then proceeded to relate that there is a third structure of the same kind in the neighbourhood," but that he had neither been able to discover its situation nor obtain any account of it." The existence of this structure he knew from Brereton but had he read his statements more closely he would have learned that it was no further away than "the next field from the St. Nicholas chamber! It must, indeed, have been among the structures he next described-" There are others, whose present state proves that they must either have fallen, or been thrown down, or else the stones must have been brought to the places where they appear, for the purpose of erecting such edifices."

Malkin belonged to the old school of antiquaries who saw in rude stone monuments the vestiges of Druidism. Possibly these cromlechs," he suggested, "were places of shelter for the Druids unless we may suppose them to have been oratories, from the tops of which they delivered their discourses, or altars on which victims were offered." That the vale was one of peculiar sanctity he found confirmed by its name, Dyffryn Golych, which he translated "the vale of worship, adoration, or prayer." His explanation of the condition of the remains was ingenious. The early Christians threw them down as objects of pagan superstition and idolatry;" but in lieu of this they threw a heap of stones" over the two largest, because they were too unwieldy to be overturned," and (now quoting 'lolo Morgan wg,' whose opinion certainly clashed with his own) "by way of showing their detestation converted them into dog or bitch kennels.

During the previous year (1806) appeared that extraordinary work, Rev. Edward Davies' Mythology of the Druids," in which one of our two chambers is referred to as Gwâl-y-filast (p. 397). His speculations are a tangle of guessing etymology and vain imaginings, in which the monuments of the class figure as places where the mystic rites of Ceridwen, "the Arkite Goddess of the Druids," whom he equates with Ceres, were enacted.

There is a short account of the St. Nicholas chamber in Carlisle's Topographical Dictionary of the Dominion of Wales," 1811, and a longer one in Lewis's "Topographical Dictionary of Wales," 1833, 1840, and 1843. They both follow Malkin, except that the width of the capstone, as given in the former, is his minimum measurement of 10 ft.

Two rare etchings on a single sheet, and signed, 'F. A. Grey, 1827,' are the earliest views of the two chambers I know of. They are very sketchy, and are reversed. The sheet belongs to Mr. William Clarke, of Wenvoe, who has kindly allowed the St. Nicholas view to be reproduced here, Fig. 5. If held before a looking-glass the reflection will show it the right way about. There is a model of our chamber in the British Museum, which was made for Sir Richard Colt Hoare in 1832, by R. Tongue. Mr. Wilfrid J. Hemp has examined it for me, and reports that it is altogether too incorrect to be worth photographing.

The first excursion during the meeting of the Cambrian Archæological Association at Cardiff in 1849, included a visit to "our aboriginal and Druidical remains the cromlechs of St. Nicholas." The visit was eventful, for it led to an extraordinary, but now almost forgotten, controversy on the Cromlech Question.' In anticipation of the visit, one of the secretaries of the Association, the Rev. H. Longueville Jones, examined the larger chamber (Arch. Camb., 1849, p. 327). He found under it three recesses branching off from the main chamber, one of which was still lined with slabs of stone. They had evidently been made to contain bodies, and the earth within them was of a dark colour, such as would be made by the decomposition of animal remains. In one to the north-east he found part of a human lower jaw, with one of the teeth in it in good preservation." Although not stated, it is evident that he did a little digging. It is also evident that he penetrated into the slab-lined entrance passage at the north-east corner, discovered during the recent excavations. It is difficult to imagine where the other two recesses were, but perhaps the gap in the north-east corner between the north and west sideslabs of the chamber was one. The chief point of interest, however, is that the sepulchral character of the remains was recognised.

On the occasion of the visit, a discussion on the spot revealed a marked divergence of opinion between the Welsh and the Irish antiquaries present, which only widened at the evening meetings. Dr. J. Henthorn Todd, F.S.A., the eminent Irish scholar and professor of Hebrew in the Dublin University, who was the chief spokesman on the Irish side, claimed that the dolmens were burial chambers, that many of the Irish popular names for them implied their sepulchral character, and that in some of them actual human remains had been found. He further pointed out that the term cromlech was unknown in Ireland, and he suspected that in Wales it was comparatively modern, at all events as a designation for these monuments. He suggested that his Welsh friends would do good service if they ascertained when and how the term came into use, and what were the old names for the structures among the peasantry. That the enunciation of the Irish view was received with some amazement is reflected in the editorial comments of the Cardiff and Merthyr Guardian," September Ist:­-''The ancient badge of our nation, upon which our bards tell us the whole of our mythology hangs, must, after the facts adduced by Dr. Todd and others, undergo a most revolting wrench: we allude to the Cromlech, the very pediment of bardism. While we write, we tremble Hitherto all our bards have maintained, with one exception, that the Cromlech is an altar and many and various are the theories founded on such a hypothesis. But shades of Dr. Owen Pugh, lolo Morgan wg and Carnhuanawc, what say you ? Dr. Todd says they are tombs, charnel-houses, ovens to bake human bones in.

After the dispersal of the visitors, the columns of that paper became the arena of the controversy, and by its close, fourteen months later, fifteen correspondents had entered the lists, the most persistent of these being Archdeacon Williams, the author of "Gomer "and other works on Welsh subjects. Most of the earlier letters were fairly within the limits of the proposed inquiry. The earliest known literary use of cromlech to designate a megalithic chamber was stated to be in George Owen's History of Pembrokeshire," about the close of the sixteenth century. While the meaning of the second syllable of the word was obvious, several opinions were expressed as to the first. It meant bowing, bent or hunchbacked, a roof or covering, the dome of heaven, a Celtic Jupiter Tonans, worship or adoration, etc. Several of the writers suggested that these structures were both altars and tombs. Only two examples of the popular names were given, and they are of special interest to us-the St. Nicholas chamber was known as Castell Carreg,' and that of St. Lythans as 'Gwâl-y-filiast.' In his replies, Dr. Todd restated his original proposition in greater detail, and hinted that the Welsh did not exactly know what cromlech meant.

On October 6th, the Archdeacon entered the lists, and in his earlier letters made considerable parade of his scholastic attainments, with many reminders of his peculiar fitness for the fray. In his opening paragraph he considered that Dr. Todd had ventured rather boldly into a field well known to me, but apparently an undiscovered country to him In his sixth letter he thought that he had "brought forward sufficient evidences to induce Dr. Todd and our antiquarian friends in Ireland to reconsider their theory," and he rattled his quiver to show that there were still more arrows in it. In his seventh, he rejoiced that the correspondence afforded him "the opportunity of discharging the heavy load of antiquarian knowledge which has so long been pent up within my own breast." In his twelfth, he was "confident that his letters, though published in a provincial paper will be a lasting record, as long as the Cymrian people and the Cymrian language continue to be living witnesses of the great facts thus closely connected with the primaeval history of the human race." The letters, however, had been of a discursive nature with very little to the point, and Dr. Todd found it necessary to restate his proposition, and he noted that "hitherto he (the Archdeacon) has been only clearing the ground by setting forth his own qualifications for the gigantic task."

By December 22nd the Archdeacon had the field to himself, and his remaining letters appeared after varying intervals till November 23rd, 1850. In his extraordinary series of seventeen letters (one filled five columns of small type of unusual width for a newspaper) and three long addenda, he dipped into the lore of all manner of rude stone monuments, which lie regarded as the works of a homogeneous race"-the Celts and ransacked the wide range of Biblical, classical, early British and bardic literature for allusions to them. His reader is lost in a maze of philological and mythological speculations, and looks in vain for any coherent theory of the dolmens-sometimes they are altars, sometimes temples, and New Grange in Ireland figures as a nymphseum. Each letter seems to be preparatory to the discussion of the main question and this, indeed, was the case, for in an addendum to his fifteenth, he stated that the most important communication was not yet ready-I; the contents of it are so very startling that I must be cautious in the wording of it! This letter never appeared, for his sixteenth was in answer to an enquirer who wished to study the Hyperboraean Question '-the subject of one of his previous letters while the last had reference to an essay on the 'Cromlech Question' at the Rhuddlan Eisteddfod, to which he had awarded the prize.

It is certain that this controversy attracted considerable notice in Wales, for not only are there many references to it in Archaeologia Cambrensis and elsewhere during the following twenty years, but it gave an impetus to the critical study of Welsh prehistoric remains. It is surprising, however, that not one of the correspondents even suggested that perhaps the spade might throw some iight on the subject, or that these monuments might be so ancient that the oldest surviving traditions and literary statements are but expressions of the opinions of times when their history and use had long been forgotten.

In Cliffe's Book of South Wales," 1847, under the head of Druidical Antiquities,' there is a short account of the St. Nicholas and St. Lvthans chambers, derived mainly from Malkin's. The crack in the capstone of the former is mentioned for the first time, and is shown in the woodcut which accompanies the text. In the edition of 1854, Longueville Jones's discovery is specified as that of "human bones."

In 1856, appeared in Arch. Camb. (p. 99) an important paper On the Names of Cromlechau,' by Thomas Stephens, the author of The Literature of the Kymri," in which, in response to Dr. Todd's enquiry in 1849, lie treated of sundry popular names of these structures and their significance but he was careful to observe that the 'Cromlech Question' could only be determined by excavation. Inter alia, he referred to one of our chambers as Llech-y-filast, and noted that Golych occurs in Liber Landavensis." We shall return to this paper later.

Another paper, 'The Parishes of St. Nicholas and St. Lythans,' by G. T. C[lark] and R. 0. J[ones], in the same journal for 1862 (p. 102) [Note. A mistake for Page 92?] is also important. It includes short descriptions of the two chambers based upon personal observation. The St. Nicholas capstone is stated to be 22 ft. by 15 ft., and its fractured condition is mentioned also that a human skull is said to have been recently found by the Rev. H. L. Jones in the chamber. The two piles of stones, C and D, are regarded as the "wrecks of smaller cromlechs," and the Quarry as not only the source of the great slabs, but as probably governing the occurrence of the group of remains. The formidable controversy between Dr. Todd and Archdeacon Williams is also referred to.

The following year, 1875, appeared in the same journal (p. 171), a paper On the St. Lythans and St. Nicholas Cromlechs and other Remains near Cardiff,' with plans of both chambers and a view of the latter, all from the pen of Mr. J. W. Lukis. The dimensions of all the stones are given on the plans, and, with few exceptions, they are as correct as possible. The chief exception is the statement that the width of the west supporting slab at St. Nicholas is 11 ft. 8 in., but this is probably a clerical error for 8 ft. 11 in., which is much nearer the truth. His dimensions of the capstone are 22 ft. 9 in. by 15 ft. 8 in., and a maximum thickness of 3 ft. 6 in. The view (a woodcut) is taken from the south-east, and as it exaggerates the height, it makes the structure look more imposing than it really is. It shows the crack very clearly, but without any displacement of the component pieces. He had no doubt that the south side was formerly closed in like the other sides, and he noted the remains of the original mound. We grubbed," he wrote, "about the debris of stones, etc., outside the cromlech and soon discovered fragments of human teeth and unburnt bones, with portions of rude pottery." He questioned some children, and understood them to call the remains, Castell Corrig,' and this he connects with the Breton corrig, a fairy, a name frequently associated with the Breton dolmens. There is no doubt, however, that lie misunderstood them, and that they really said Castell Carreg.' Curiously, some other children called the St. Lythans chamber, Stony Castle', and this should have suggested the probability of a misunderstanding with regard to the other name.

A paper on The St. Nicholas Cromlechs,' by Mr. Franklen G. Evans, F.M.S., in the "Transactions of the Cardiff Naturalists' Society," 1881, is really a dissertation on megalithic monuments, Druidism, etc., with only two or three references to our chambered tumuli. He regarded such ancient monuments as post-Roman, because "the Romans do not mention these huge structures." Incidentally, he considered that Lukis was "mistaken about Castell Korrig," and that what he was told was Castell Cerig."

An enquiry from Mr. Octavius Morgan in the "Western Mail," on May 12th, 1882, for information whether there was in Glamorganshire any cromlech or large stone called Gwal y Filast, or any similar name," set the ball rolling for a few days' correspondence. Among the replies was one from Iorwerth,' who "could not find mention of any Gwal-y-Filast, but there is a Llech-y-Filast at DuffVyn, St. Nicholas, near Cardiff, which is considered to be the largest cromlech in Britain, measuring 17 ft. by 13 ft., the height not mentioned, but I believe from 8 ft. to 10 ft." It is not clear as to which of our two chambers he referred, as the measurements agree with neither. In the same issue of the paper was a letter from in which he gave the locale of Gwal Llech-y-filast" as three-quarters of a mile to the south-east of Dyffryn House, and this would be the St. Lythans chamber. The Rev. (now Archdeacon) Conybeare Bruce, who wrote with local knowledge, understood by Llech-y-Filast, the St. Nicholas chamber, and corrected Iorwerth's measurements by Malkin's. He added, inter alia, "I I am sorry to say that a crack has appeared in the roof-stone, towards the lower and narrow end, and I fear, unless it be propped up from the inside, the roof-stone may ere long fall in." It is noteworthy that in this correspondence there is not a hint of a possible sepulchral origin for the remains. On the contrary, several of the writers enlarged on their supposed Druidical connection in extravagant terms. To the Vale of Worship" resorted the Druids to adore the Eternal Celi" and the St. Nicholas chamber was a grand Druidical Cathedral!

In 1883 it was proposed that "the true origin and purpose" of the megalithic remains should be a special subject at the forthcoming meeting of the Association at Fishguard. To this the Rev. E. L. Barn well objected on the ground that the question had already been settled years ago and it led to a paper from his pen, On some South Wales Cromlechs,' in Arch. Camb. the following year (p. 129). In the course of this paper he pointed out that all the older antiquaries from Camden to Fenton had said that the cromlechs were nothing more nor less than Druidic altars," but that none had brought forward any reasons-even remote traditions-in favour of their theories. He, on the contrary, claimed that they were burial receptacles, and were originally covered with mounds. Among those which still retain remains of their mounds, he instanced our St. Nicholas (under the name of Dyffryn) chamber.

The British Association visited our remains during their meeting at Cardiff in 1891. In the Programme of Excursions they are designated dolmens,' a term very rarely used in Wales, and the description is mainly drawn from Barnwell's of 1874, but the mistake is made of applying his dimensions of the St. Nicholas interior, 19 ft. 9 in. by 11 ft., to its capstone.

On October 29th, 1898, the late Mr. John Storrie wrote to the Western Mail" that he knew of three perfect and nineteen broken or imperfect cromlechs near Cardiff," and that he considered that they were erected not earlier than about A.D. 500 or 600. If they had been standing," he went on to say, "in the time of the Roman Occupation, it is incredible but that some trace of Roman adaptation would be found on one or other of them, or that some 'Bill Stubbs' might have left his mark somewhere about them, or some reference to them might occur in the Roman literature relating to these parts." He also mentioned that he was present at a meeting of the Cardiff Naturalists' Society at the Tinkins Wood Cromlech when Mr. Lukis opened h small chamber in the north-east corner and found the residue of an interment there." This statement is puzzling. The society made a passing call on their way to St. Donats on July 30th, 1874, when Mr. Lukis addressed them but he said nothing of this small chamber or of any discovery of an interment, and there certainly was not sufficient time for any digging on that occasion. In his communication to Arch. Camb. the following year he did mention his finding of human bones, but they were "amongst the debris of stones, etc., outside the cromlech." Probably Mr. Storrie was with Mr. Lukis when he found these, and, writing more than twenty years later, unwittingly mixed up this discovery with Mr. Longueville Jones' of 1849. The next visit of the Naturalists was in 1881, some years after Mr. Lukis had left South Wales for good.

In the course of an interview with a representative of the "South Wales Evening Echo" (September 22nd, 1899), I expressed my conviction that these structures were burial chambers, and that so far from being Druids' altars, they were raised long before Druids were thought of. In illustration of our chambered mound in its original condition, I made a sketch-plan, which was reproduced in the paper.

On August 22nd, 1907, Sir Norman Lockyer, F.R.S., delivered a lecture at Swansea on The Age of the Gorsedd," as determined by astronomical data, and this led to an article from his pen in "Nature," on December 19th following, in which he recorded, inter alia, a visit to our remains. It is clear from his lecture that he regarded the structures known as cromlechs in Wales as chambers or receptacles originally covered with mounds. While this disposes of the altar theory, he did not admit them to be tombs. "They were never meant for tombs, and to say that these things were built as sepulchres, just because people have since been buried in them, is to deny that a church was built for the worship of God because you find corpses in it." He regarded them as "priests' houses." The covering mound was to keep out the rain, because they had their fires to look after, not merely for cooking purposes, but for sacred purposes. Then these people wanted water to drink, and it was only natural that they should have a water-supply in these cromlechs. Might not that be the real origin of much of the crockery that one finds in these places ? In his Nature article lie described the St. Nicholas structure as "a large rectangular cromlech in Duffryn Golych or Goluch (the Vale of Worship) and called by the natives Castell Corrig (Dwarf's Castle), a name which suggests belief in the presence of fairies there." He was not able to determine the exact astronomical bearings without some excavation but from a letter to the "Western Mail" of September 5th, from his Welsh colleague, the Rev. John Griffith, we learn that he tentatively assigned the date B.C. 3000 to the remains.

During the visit of the Cambrian Archaeological Association to this structure on July 24th, 1912, the President, Dr. Boyd Dawkins, F.R.S., expressed his opinion that it belonged to the Bronze Age, and he dissented from Sir Norman's views on the ground that prehistoric man had no knowledge of astronomy." In the short report of the visit in Arch. Camb., 1913 (p. 100), the capstone is said to be 25 ft. by 16 ft., that is 1 ft. greater in length than Malkin's measurement. Mention is made of the remains of a large circle," and of blocks of stone "that may be parts of an avenue," also a cistfaen in the adjoining field. So far these have escaped my scrutiny. In the description of the remains in Mrs. M. Robertson Spencer's "Annals of South Glamorgan (p. 174), the dimensions of the capstone are given as 28 ft. by 15 ft. Considering that Brereton's in 1772 were 18 ft. by 15 ft., who can now doubt that stones grow!

During the last twenty-five years, the remains have been noticed in several guide-books to South Wales and in several educational works relating to Glamorgan. In the former they are usually regarded as Druidicalin Marie Trevelyan's Holiday Haunts in Glamorganshire," 1899, for instance, the region is said to be "the great Druidic centre of Siluria, in the historic Vale of Worship "-whereas in the latter they are invariably held to be sepulchral. Twice the St. Nicholas chamber is said to be known as Llech-y-filast and once as Castell Corrig.

In perusing these literary gleanings the reader will have observed that our chamber and its companion of St. Lythans have been the subject of a diversity of opinions as to their purpose and age, and incidentally that these opinions represent how the rude stone monuments of their class have been regarded the last century-and-a-half. According to the earlier writers quoted, they were Druidical altars or platforms for priestly orations or religious exercises-hence were structures intended to be exposed and seen whereas, according to most of the later writers, they were burial vaults hidden within mounds, and, if now exposed, this is due to the removal of the envelopes. The latter is the modern view and the persistence of the older is to be set down to ignorance of the trend of modern research. Almost invariably those which have been subjected to the spade have yielded evidence of interments and around many which were supposed to have always been 'free-standing cromlechs' have been found the traces of their covering mounds. Malkin noted the remains of the mounds at St. Nicholas and St. Lythans, and his theory compelled him to regard them as anti-Druidical additions of an after age, and not as essential features of their original state.

Sir Norman Lockyer stands somewhat apart from either school. He admits that the structures were chambers and were formerly covered with mounds. He admits also that at least many were used for interments but he denies that this was their original use-they were built, he asserts, for priests' houses. His analogy of burials in a church is scarcely a parallel one, for these burials do not interfere with the carrying out of public worship whereas the introduction of the dead into a dolmen would render it untenable as a habitation for the living. This introduces questions which he seems to have overlooked. Why should the priests have vacated their abodes ? Are we to understand that the burials took place in a subsequent age when these structures were derelicts of a forgotten religious system ? His theory to be consistent should be of general application. We can imagine that large structures, as those of St. Nicholas and St. Lythans, might have served as habitations but it is incredible that the smaller were 30 used, as, for instance, one I excavated on Harborough Rocks in Derbyshire, the chamber of which had a mean length of barely 3 ft., a width of 2 ft., and a height of 2 ft. 4 in., while its access gallery, to judge from its scanty remains, could scarcely have exceeded 18 in. in height. Even if the tenant was a dwarf, there would not have been space enough for his fire and his water-pots as well as himself. And what about the smoke from the tire? Does Sir Norman include chimneys in his theory? Another question -how came it that while the laity enjoyed commodious circular huts with thatched roofs, the priests, who should have been highly honoured, were consigned to dark and damp underground cells ?

The age of the chambered tumuli is a difficult problem. The eighteenth-century antiquaries appear to have been unanimous in assigning them to a remote antiquity and in spite of Fergusson's arguments in his "Rude Stone Monuments," 1872, to prove their comparative modern origin (which at the time obtained a considerable following), the consensus of present opinion is strongly in favour of the old view-that is, that they are prehistoric. The prevailing view is that they belong to the Neolithic period, but some antiquaries - Professor Boyd Dawkins for one - regard them as of the Bronze period. This, however, is too intricate a question to be discussed here, especially as our recent excavations at St. Nicholas have failed to throw light upon it. One of Fergusson's arguments is echoed by Mr. Storrie in his Western Mail letter of 1898-tiaiiiely, that if these great structures were in existence during the Roman period the contemporary writers would have mentioned them. So far as Britain is concerned, however, the literature of the period is by no means extensive, and is practically confined to geographical and military lists and narratives of the doings of the governors and their armies-a literature not likely to notice the rude stone monuments of the natives, the meaning and use of which were probably even then forgotten.

We now turn to the folk-lore of our subject. The present popular name, The Cromlech, may be dismissed as one of recent introduction. Castell Corrig has already been ruled out as a recent blunder for Castell Carreg, but while this may be an old name, it is obviously a fanciful descriptive one, and so far as is known is not associated with any legend. There remain Llech-y-filast, equally shared by the St. Lythans chamber, and Gwal-y-filast, a name of the second, and probably also common to the two. There is reason for thinking that these names are ancient, and they appeal to the imagination as having their source in some old-world myth. According to Malkin, Gwal-y-filast was in his day a frequent designation of these structures in Glamorgan, but this is not the case now. This and kindred names, however, are widely distributed in South Wales, and appear to be very rare elsewhere in the Principality.

In Thomas Stephens paper on the Names of Cromlechau,' already referred to, the following examples are given — Llech-yr-ast, Cardiganshire; Gwal-y-filast, Carmarthenshire and Monmouthshire¹; Llech-y-filast (two), Glamorgan Carnedd-y-filiast, Denbighshire and Llech-y-fleiddast (stone of the she-wolf), but the locality of which he had forgotten. To this list we add the St. Lythans Gwal-y-filast, and another in the parish of Llanedern², Glamorgan, mentioned by Gibson in his additions to Camden's Britannia" (1695 ed., p. 646) also a Twlc-y-filast at Ffynnon Newydd, Carmarthenshire (Arch. Camb., 1876, p. 236).

With regard to the prevalence of these names in South Wales, Stephens pointed out that the distribution corresponds with the line of the country traversed by King Arthur in his legendary hunt of the Twrch Trwyth," which is so graphically narrated in the story of Kilhwch and Olwen.' This suggests some connection between the remains which bear these names and Gast Rymhi or Rymi³ but to quote Stephens Gast Rhymhi was originally a female, and most probably one of distinction and for some reason or other, in accordance with mediaeval ideas, she was transformed into a she-wolf. It is at this stage that she is noticed in the story. Kilhwch demands the assistance of Arthur and his knights to obtain the hand of Olwen. his lady-love and, among other things, he demands the two cubs of Gast Rhymhi.' Arthur goes in search of this animal, and enquires where she is some one answers that she is at Milford. The hero found her in a cave at that place and having surrounded her and her two cubs, the story states that God did change them again for Arthur into their true form.'

Note 1. 1 A remnant of a chamber on the east side of the Rhymney Valley, near Michaelston-Fedwy, and now known as Druidstone.

Note 2. This parish faces Michaelston-Fedwy, on the opposite side of the river. Perhaps Gibson intended the Michaelston-Fedwy Gwal-y-filast.

Note 3. So spelled in the original.

The circumstances of the discovery do not imply that this cave was her home, but rather that she was a cave-haunting creature. We can well imagine that popular fancy would associate her with any cave or cave-like hole in the region of the hunt, such as, for instance, the St. Nicholas chamber, especially when covered with its mound. Her name suggests a connection with the Rhymney, and Stephens noted the presence of a Gwâl-y-filast near this river as singular, but he regarded it as a coincidence. Considering that there was probably another close by, and that the St. Nicholas and St. Lythans Llech-y-filast and Gwal-y-filast are within a few miles, the association seems to too remarkable to be a coincidence.¹ Mr. Ifano Jones, the Librarian of the Welsh Department. Cardiff, informs me that in the old Gwentian dialect the n would be dropped in Rhymney (or Rumney), and that it is still often pronounced Rummy.' I consulted Sir John Rhys, and he replied that "in the Liber Landavensis it is a syllable longer than it is now­-namely, 'Gurrimi,' Guorrimi,' the prefix being unaccentuated has dropped-hence' Rymi,' not Rhymi, for, as far as I can remember, I have never heard it called the latter. Mr. Jones is probably right that an n has been dropped, so that the name would be originally Guorriuun.' The Cambro-British Saints (p. 45) has 'Rymni' and Remni,' but the MS. has 'Rymi' twice -the editor of that volume has done his work carelessly. It is possible that the dropping of the n in some of these instances was a mere accident of copying, that consonant being perhaps represented by a dash over the m (Rymi). Butthe 'Kilhwch and Olwen' has Rymi' and Rymhi,' and the h may be an error for n -that is, an error for Rymni.' This, however, is a mere guess."

Note 1. See also Wirt Sikes' "British Goblins," p. 380.

Sir John refers, not to the Monmouthshire Rhymney, but to a stream which, in the Liber, is associated with the Golych as forming part of the boundary of a tract of land in the vicinity of St. Lythans, granted by King Iudhail to Llandaflf. In the paragraph in the Cambro-British Saints,' referred to above, the two streams are again associated, the Golych as the eastern and the Gurimi (as it is spelled) as the southern limit of the lands of St. Cadoc, in this part of Glamorgan;

Continues ...

Some Characteristics Of Chambered Tumuli.

My first visit-now more than twenty years ago-to the St. Nicholas and St. Lythans chambered tumuli, convinced me that they are of a type which prevails in Wiltshire and Gloucestershire, but is by no means confilled to these counties. For the reader to whom the subject is new, a summary of the characteristics of this type will help him to follow the details of the excavations which will be given in the next section. In this type the mound is long, and wider at one end than at the other, and the wider end usually faces the east or approximately so. In every instance in which the mound has been excavated, the wider end was found to be developed into two more or less rounded lobes or 'horns,' the intervening indent being somewhat funnelshaped. Normally the indent was the approach to the chamber, or the set of chambers, which in either case was axial-that is, was placed on the central longitudinal line of the mound. The entrance was, of course, at the head of the indent, and it sometimes had the form of an imposing portal, vying with, or even surpassing, the receptacle in its megalithic character. In several instances the stone slab which served as a door has been found. Between the portal and the receptacle there was, probably invariably, a tunnel-like avenue or gallery, long or short, and from 2 ft. to 3 ft. or more in width, and about as much or a little more in height. The receptacle was often a single terminal chamber but more often the gallery was prolonged and had lateral recesses, and as these usually faced one another, each pair may be regarded as a transeptal chamber. In either case the receptacle with its gallery, was erected on the old natural surface of the ground.

In an important variant of the type, the receptacle or receptacles were not axial, but lateral, and each was entered from the contiguous side of the mound. In these mounds the indent at the wider end was retained, but it was a blind one, and the place occupied by the portal in the normal type was marked by a slab or some structure which simulated an entrance, and clearly was a survival of it. The mounds appear to have been always constructed of stones, but occasionally a large admixture of earth, probably derived from a capping of the same, has been reported. A prominent feature was an external revetment or 'podium,' usually built of thin undressed stones, neatly fitted together, and forming what a mason would describe as random work. Occasionally this enclosing work consisted of a series of large upright stones, with the intervening spaces filled in with walling of the above character. The remains of the retaining-wall are rarely found to exceed a height of 3 ft., except along the sides of the indent, where, at the head, a height of 4 ft. to 5 ft. is not unusual. That the wall should be preserved to these heights is due to the talus or accumulation of debris derived from the mound and the wall itself, which acts as a support and a protection, its better preservation in the indent being due to the greater height of the debris there in consequence of the confined space. The mound itself was often raised with considerable care and method, the stones being placed inclined on edge or laid flat. Occasionally, internal lines of rough walling have been observed, as if erected with the view of dividing the space to be occupied by the mound into compartments.

A small chambered tumulus which was excavated at Park Cwm in 1869 (Arch. Camb., 1871, p. 168) under the supervision of the late Lord Avebury, shows that the distribution of the Wilts-Gloucestershire type extended westwards at least as far as Gower. It was 60 ft. long and 50 ft. wide, with the entrance to the south; and its gallery threaded two transeptal chambers, in which were found the fragmentary and scattered remains of twenty-four individuals. Further west, and especially in Pembrokeshire, there are many denuded chambers, but nothing is known of their mound arrangements. In mid-Wales they are few but in several parts of North Wales, especially in Carnarvonshire and Anglesey, they are numerous. Very few of these have been subjected to the spade two, however, the one at Capel Garmon, Denbighshire, and the other at Plas Newydd, Anglesey, have been, and have given results which show that they differed substantially from the above type. Their mounds were oval chambers central and these were entered, not from one end, but from one side of their respective mounds (see plans, Arch. Camb., 1856, p. 91, and 1870, p. 51). Considering the large number of these remains in the Principality, the Welsh archaeologists cannot be congratulated on the amount of scientific research they have bestowed upon them!

The Excavations And The Structural Remains.

In the autumn of 1907, Mr. William Clarke and myself examined the St. Nicholas chamber, and were convinced that the displacement of the larger portion of the capstone had considerably increased of late years. I subsequently wrote to the late Mr. John Cory, the owner of the site, recommending the insertion of a pillar, also that the chamber should be cleared of its rubbish and a few trenches cut to ascertain the nature and limits of the mound. The proposal was entertained, but it was too late in the season to be carried out. The subsequent resumption of the Gelly-gaer excavations necessitated further postponement, and with Mr. Cory's death the matter dropped. A year or two later, the Rev. Prebendary Hamlet, a frequent visitor at the Dyffryn, revived the project, and not only obtained Miss Cory's consent, but also her kind offer to supply the labourers and any material that might be required. Under these favourable conditions the exploration was carried out with a thoroughness that would otherwise have been hardly possible, especially as the work was of a more arduous nature than was anticipated. The opportunity was also seized to remove the surrounding talus of debris so as to leave what remained of the mound in clear relief and to make certain restorations which will presently be described. From the outset, Mr. Reginald Cory was keenly interested in the work so also were the Council of the National Museum of Wales and the Director, Dr. W. Evans Hoyle, M.A.. from whom I received every facility for superintending the operations. Mr. Clarke also rendered good service, especially with regard to the necessary steps for insuring the safety of the capstone, for the carrying out of which he supplied two skilled men.

Actual operations began on June 8th, 1914, with three men, but were suspended on the 16th in consequence of pressure of work on the estate. They were resumed on August 17th, and went on continuously till October 22nd, the number of men employed varying from four to six, according as they could be spared. During these eleven weeks the weather was ideal, and the men took an intelligent interest in the work and were all that could be desired.

The main object at first was to ascertain if the mound had a retaining-wall, and, if found, to follow it up. Choosing a spot on the north side and well beyond the limits of the mound, we cautiously cut a broad trench towards and into its slope, and at the end of about 16 ft. found the wall, here reduced to two courses and much displaced. It was a comparatively easy matter, now that the line was known, to expose it at several points along the north side. All these trenches showed that the slope exterior to the wall was a pell-mell accumulation of debris, amongst which were many tabular stones that had been used in its construction. The next step was to find the revetment on the south side, and this proved to be a tedious operation. The first cutting was a little to the west of the middle of that side. After passing through the talus of debris, something more orderly was entered-great rough stones closely packed together, with the interstices filled with tough clayey soil. This was our first insight into the construction of the mound. No wall was found here, but the thin tabular stones in the talus were a hint that we had passed its line. The next search was made at the extreme east of this side, but after considerable digging the only evidence of the wall was its scattered stones. A third and intermediate trench on the eve of the suspension of the work was crowned with success.

On the resumption of the excavations on August 17th, the men were set to complete the exposure of the revetment, but on both sides it was entirely lost towards the west end of the mound. As stated on p. 257, the western limit of the mound was obscure, and the only feasible procedure was to cut a long central trench, working from west to east. This brought to light the revetment at a point more eastward than was anticipated, and eventually a long stretch was exposed, but straight, not curved as expected. All our efforts to join up this with the north and south revetments failed the corners of the mound were entirely gone. The north revetment had been followed up eastwards to 2 ft. beyond the fence between the plantation and the field, where it was found to end abruptly on the west side of the ditch. No surface-indication of the mound was visible in the field, and we concluded that its eastern portion had been removed. In the hope of finding some traces, we cut several trenches in the field, but without any success. On the east side of the ditch, however, a vague line of flat stones, starting from a point opposite the abrupt end of the north revetment and extending southwards about 14 ft., was traced. It certainly looked like a right-angled return of that revetment, but it was equally likely to be an edging to the ditch. All our endeavours to ascertain the original eastern limit of the mound had failed so far.

The next question to be solved was the original means of access to the chamber. The normal position would be from the nearer end of the mound, which in this case would be the east end but this side of the chamber appeared at the time to be wholly closed in, like its north and west sides. The open south side suggested possibilities, and there were known instances of a chamber being entered from the side of a long barrow. The fact that the chamber lacked a south wall, suggested a work of destruction which might very well have involved that of an access-gallery and the fact that the mound on this side was very much reduced certainly favoured the hypothesis. So we made an appeal to that grand resolver of archaeological doubts-the spade. The remains of the mound here were carefully removed in the hope that at least the site of a gallery would be revealed but nothing of the sort was brought to light. The digging, however, was not in vain, for amongst the superficial debris in the immediate vicinity of the chamber, a large number of broken and scattered human bones were found, many extremely decayed, and in addition a few animals' bones and fragments of rude hand shaped vessels.

It was not deemed advisable to thoroughly excavate the interior of the chamber until its broken roof was properly supported but it was necessary to clear a space down to the solid ground for the insertion of the proposed pillar. In doing this much of the superficial debris throughout the interior was removed, and after the insertion of the pier the clearance was completed. The accumulation varied in thickness, but the average was about 2 ft. It exhibited no definite stratification. At the surface it consisted of stones in a loose blackish leaf-mould, with a few animals' bones, and here and there a piece of modern earthenware or glass bottle. Below this, the mould was more compact and earthy, and not so dark and this passed downwards into a mottled brownish matrix in which the natural soil of the site predominated, and sufficiently stiff and tenacious to require the constant use of the pick to dislodge the stones. This rested on the floor of the chamber, and interspersed in it was an abundance of broken human bones, with a few fragments of rude pottery, some of hand-made and others of hard wheel-made vessels. There was no evidence that the floor had been paved, as these chambers often were and the actual floor-surface was not detected, but it was certain that the builders had pared off the superficial turf and mould down to the hard marl. In several places at the foot of the north and east sides were patches of reddish clay; and in two of them on the former side were embedded parts of the calvaria of two human skulls, but so hard and tenacious was the clay that we were only partially successful in removing them. There was no evidence that this clay had overspread the floor on the contrary, it had the appearance of having been washed in by the percolation of rain-water.

We were convinced of two things — (1) That our excavation in front of the chamber was in the normal mound-structure or cairn, and (2) that the accumulation within was adventitious, consisting of stones and earth, which at first had been thrown in or had rapidly rolled in, and which later had been so slowly introduced as to allow of the predominance of dark mould, derived from leaves and other vegetable matters blown in by the wind. We were also convinced of something else. Instead of the external cairn tailing off or merging into this accumulation, it ended abruptly at the south side of the chamber, and this admits of only one explanation-that this south side was originally closed in like the others. From these data we can, I think, reconstruct an episode in the history of our chambered mound. Reasons have already been given for believing that sometime long ago the mound was used as a quarry, and that the stone-gatherers attacked it from the south. If at that time the mound was fairly intact (as is probable), and that the chamber was buried in it, with or without the capstone being visible, the men would become acquainted with its existence by reaching its south side. The discovery would arouse their curiosity, but it would not be the curiosity of the archaeologist. We know that all over the country prehistoric burial-places have been broken into by treasure-seekers, even in quite recent times. Prompted by the possibilities of gain, our men would reduce the mound in front sufficiently low to enable them to break into the interior. In their search they would throw out some of the contents, and rummage amongst what remained. This exactly explains all we observed-the absence of a south side or wall, the great reduction of the mound in its vicinity, the confusion within, and the scattered bones without.

I now narrate a most important sequel to this phase of the exploration. Previously the three great stones on the east side of the chamber appeared to completely close in that side but now that the accumulation was removed and these stones fully exposed, the northern one was seen to have some resemblance to the letter L upside down, its upper end being prolonged southwards and leaving the broad gap below, indicated by F in the low-level plan, Fig. 12, and the elevation, Fig. 14.¹ Before leaving for a short holiday (during which the masons inserted their pier), I gave instructions for the hard soil which filled it to be removed, and, if it passed through the side of the chamber, for a hole to be sunk down to it on the outside. The men found that it not only passed through, but was continued as a passage (G on the plan) between the two thin slabs. On my return, I had the felicity of congratulating them on the discovery of the entrance to the chamber.

Note 1. There is little doubt that this block has fallen considerably forwards in the direction indicated by the arrow in Fig. 14, and thus has disfigured the shape and reduced the height of the aperture.

A week's work disclosed the whole arrangement of the east end of the mound. After removing several cart-loads of debris, the spacious approach, H, Figs. 10 and 12, was cleared. This approach is of the full width of the chamber at the head, and expands eastwards between the curved revetments of the horns,' that on the north side attaining a height of 4 ft. 3 in. at its junction with that structure. In following up this revetment, we found that the line of stones (p. 296) on the east side of the field-ditch was its northern continuation and it was evident that when the ditch was made the northern horn was cut back to make room for it. Of the corresponding continuation of the opposite revetment no trace remained-it had been completely obliterated by the ditch-makers hence its line, as shown on the plan, is conjectural.

These two revetments are carried right up to the angles of the east side or front of the chamber. There was good evidence that the actual structure of the chamber was not originally visible. Along its foot stretches the remains of walling (i) several feet in height, which differs from the revetments in its pronounced batter and slighter and rougher construction. In fact, it may with propriety be regarded as a wall-like bank of stones and earth, with a pitched facing, such as one often sees as a finish to hedge-banks. If the slope of its front is projected upwards, it approximates to the front edge of the capstone, and this strongly suggests that the bank was originally carried to that height leaving the edge of the capstone exposed. The floor of the approach is of earth with much small stone, and hard as if it had been beaten; and it has a gentle fall to the east.

The slabs which form the sides or jambs of the entrance have been carefully selected in order that their front edges may conform with the slope of the bank. The width of the passage at the ground level is 2 ft. 7 in.; but as the sides slightly incline inwards (owing, no doubt, to the pressure of the bank on either side) the width above is slightly less. Both slabs are broken, and the loose upper portions were found lying amongst the debris which choked up the passage; and above them lay part of another slab which probably formed the lintel. Upon removing the debris another large slab, broken into two pieces, was found lying on or near the bottom. The sides of this slab have been slightly, but roughly, trimmed in order to make them parallel, and in spite of its irregular upper end, due perhaps to the loss of its right corner, it approximates to a regular oblong 3 ft. 1 in. by 2 ft. 6 in. This stone would exactly fit into the entrance had its sides remained upright, and there is no doubt that it was the door, which had fallen backwards into the passage. The total length of the passage is 5 ft. 5 in., and its bottom is packed with rough stones. As the floor of the approach is about a foot higher than that of the chamber, the descent into the latter may have been by an incline or by steps, the rough stones serving as a foundation in either case. Immediately in front of the external threshold the ground was slightly and very roughly paved, and at the threshold itself two larger stones were so placed as if to serve as a sheath for the foot of the door. As the jambs of the entrance have no reveals, it is probable that the lintel was so constructed as to provide a ledge for the top of the door to rest against.

The accumulation which filled the approach consisted very largely of stones such as would be used for the contiguous revetments and the pitching of the embankment at its head, and this convinced us that these structures were considerably higher than their present highest remains. From all these data, we can in some degree reconstruct the east end of this ancient mausoleum. In the centre was the spacious approach between the bold incurved revetments of the horns.' Stretching across the head of the indent was the sloping wall-like front of the chamber surmounted with the horizontal edge of the capstone, like a rude cornice, the top of which was 9 ft. above the ground while below, near the right end, was the entrance with its slab door. How high the revetments of the horns were we cannot say, nor yet whether the top of the capstone was covered with the mound or was exposed to the sky.

It should be mentioned here that a few human bones were found in the passage and on the floor of the approach in its vicinity, where also occurred a considerable quantity of coarse black potsherds extremely decayed, and apparently belonging to one vessel.

The mound now claims attention. Almost invariably the chambered tumuli which have been explored have yielded sepulchral remains other than the original chambers. These secondary remains may be chambers themselves, cists or other enclosures, while unprotected interments are not infrequent. They belong to various periods, and indicate that the great mounds in which they are found, were resorted to for burial purposes, in some instances after intervals so long that funeral customs had entirely changed. The secondary chambers must be considered as belonging to the era of the mounds themselves. Cists, although not confined to the Bronze Age, are highly characteristic of it. Against the revetment of a chambered tumulus at Five Wells, Derbyshire [Map]¹, was found a cist containing a contracted skeleton, erected against the face of its revetment, and it indicated that at the time of its erection there was a considerable accumulation of debris. In similar tumuli at Eyford, Gloucestershire [Map]², there were skeletons associated with Bronze-Age pottery, and at Uley [Map]³, in the same county, a high-level interment associated with Roman coins. A small chambered cairn at Mininglow, Derbyshire [Map]⁴, had been augmented on one side with an earth mound containing a late Bronze-Age cremated deposit. Even the original chambers and passages were sometimes invaded by later peoples for burial purposes. These examples indicate how very important is the evidence of the secondary remains in determining the antiquity of the chambered tumuli.

Note 1. "Reliquary, 1901, p. 240.

Note 2. "British Barrows," p. 517.

Note 3. "Archæological Journal," xi, p. 315.

Note 4. Vestiges of the Antiquities of Derbyshire," p. 41.

In order to ascertain the construction of our mound and what it might contain, we adopted an expeditous system of excavation which left little chance of any cist or interment escaping detection. It consisted primarily of a series of diagonal trenches, about 3 ft. in width, carried down to the old natural surface. In the event of meeting with anything unusual that might prove to be a clue, this was followed up to a decision before resuming the diagonal in which it was noticed. In the end, the plan of the excavations presented an irregular network threaded by these long diagonals. There were, however, exceptions. The diagonals left untouched a large triangular space within the north-west corner, which was excavated as a whole and a single diagonal across each of the depressed areas, B, B, Figs. 3 and 10, was considered sufficient, as in each case what was left of the mound had been completely turned over by the stone-gatherers. In some of the longer diagonals, labour was economised by a process of refilling as the work went on. A short length of a proposed trench was first cut, and the upcast was entirely removed. After this the diggers threw the upcast into the void behind them, and so on to the end of the trench. In all the trenches small holes were cut every few yards into the natural soil to ascertain its geological nature and whether it had been disturbed. This sweeping' of the mound was a laborious operation, for most of the stones were large (some so large that they could only be moved by the use of bars) and difficult to dislodge.

As a rule, the builders deposited the stones with some degree of method. Frequently they were set on edge, usually with an inclination towards the centre.¹ Less frequently, they lay more or less flat, and it was comparatively rare that no perceptible order was observed. While as a rule the stones were of all sizes intermingled, in some places, large or small greatly preponderated. In the vicinity of the chamber they were unusually large, and leaned towards it, with the evident intention of providing a firm support for its sides.

Note 1. This arrangement has been observed in other chambered mounds, notably in those at Ablington, Eyford [Map], and Nether Swell, in Gloucestershire.

One of the earliest of the sweeping' trenches brought to light the ruined enclosure L, Fig. 10, which is stili kept open to view. Its north and much of its east and west sides are lined with thin slabs, but most of them have their upper portions broken off. Probably the same hands which wrought this mischief removed the missing slabs on the south. The interior is roughly 9 ft. 6 in. square and about 34 in. deep, the bottom being the old natural surface. There is little doubt that it is a secondary work, as the mound in its immediate vicinity had been much disturbed. Whether it had ever been covered with a large slab or had been otherwise roofed is uncertain. It has been suggested that the superficial group of cromlech-bed blocks close by its west side are the relics of its capstone, which, it is assumed, was broken up by the stone-gatherers. I could find no evidence for or against this; but I am inclined to think it more probable that they came from the hole which was made for the enclosure. If it was intended that it should have so weighty a capstone, thicker and stronger slabs would surely have been used for its sides. When discovered, the interior contained a pell-mell accumulation of stones and earth, amongst which were a few decayed bones of animals, apparently oxen and sheep but nothing was found to throw light upon the use of this enclosure. In its vicinity on the south, and especially on the west, were more decayed bones, apparently of the same animals, with a few teeth of pigs and bones of sundry small animals, all found on or near the surface. With them and close by the enclosure were several fragments of human bones-the only human remains found in our diggings in addition to those in and near the chamber, and nowhere else were animals' bones otherwise than rare.

A curious and puzzling feature of the mound will now be described. While thin rough stones were of frequent occurrence, and in all sorts of postures-flat, tilted and occasionally vertical --there were in several places in the southern half, upright slabs of considerable sizes, varying from 18 in. to 2 ft. in width. Most of these are shown on the plan, Fig. 10, and it will be observed that they are so placed that their edges are approximately north and south. It will also be observed that in several instances two or more are in line, the alignments (K, K) having a similar direction. The longest and earliest discovered of these alignments was about 18 ft. from the west revetment. The slabs were from 2 ft. to 2 ft. 6 in. in height, and reached down to the natural soil. They were not in contact, and one of the intervals was considerable. At first we regarded them as the remains of the side of a similar enclosure to that described above, but upon further investigation found that they were simply a line of slabs in the mound. This applies equally to the other upright slabs in this half of the mound.

When excavating the untouched triangular space within the north-west corner, one of the men noticed several small slabs slightly protruding through the surface of the mound, which were in a line running parallel to the west revetment at a distance of 23 ft. Further examination revealed a similar line 141 ft. to the east. The digging now had special reference to the elucidation of these lines. Each was soon found to consist of a row of stones of various shapes, often in contact and rarely widely separated. Very few of the slabs exceeded a foot in width. Some were rooted Fig. li.-Part of a Row of Stones. Revetment shown in Section into the old natural surface, and of these two or three were sufficiently tall to show on the surface of the mound, and several smaller, which attained the same elevation, were held in position by the grip of stones inclined against them on both sides, as indicated in Fig. 18. Many of the stones, on the other hand, were blocks of any shape simply resting on the floor of the mound. Eventually another row was found midway between the two. Although the rows were too obvious to admit of doubt, it is probable that, except for the fortunate discovery just mentioned and the consequent special digging, they would have been overlooked, as most of the stones which contributed to them were indistinguishable from the generality of those used in the construction of the mound. Fig. 17 is an elevation of the western of the three rows. In this all the stones, with the exception of the third and fourth from the revetment (on the right), were rough slabs, and three of these showed on the surface of the mound.

That the various rows described above were placed before the mound was raised, cannot be questioned. Had they been of subsequent introduction, trenches must have been cut through the mound for their insertion but there were no signs that such trenches had ever been made, and it is scarcely conceivable that they could have escaped detection.

With regard to our chambered tumulus as a whole, the trend of evidence points to its construction in the following order:-The first work appears to have been the chamber. Its site having been prepared by the removal of the turf, the side slabs were set up and were supported externally by a cairn of large stones in order to prevent their shifting while the capstone was being placed and to further ensure this, it is probable that the interior was filled with stones, which were afterwards removed. The second stage of the work seems to have been the revetment, at least to a sufficient height for the next stage, the mound. In this, I think we can discern three operations-the parcelling out of the void space within the framework of revetment by rows of stones, as already described, the raising of the cairn and the addition of the earth capping. Also subsequent to the revetment, and perhaps last of all, was the clothing of the front of the chamber with its sloping wall or embankment, and the construction of the outer portion of the entrance.

That the mound consisted of a cairn with a capping of earth can scarcely be doubted; but while the original height of the whole is beyond ascertainment, there is some reason for thinking that the cairn had the form of a level platform of an average height of between 3 ft. and 4 ft. The highest remaining portion before the excavations had, as stated on page 307, a flat summit, and it is noteworthy that the highest upright slabs attained this level, and none exceeded it. It impressed me as being a portion of the original summit of the cairn for if the summit had been higher it would be difficult to understand how the operations of the stone-gatherers should have left a level tract, and, still more, why this should happen to be the highest remaining portion of the mound. The thickness of the earth-capping, and whether it was flat or heaped up to form a convex summit, are matters of pure conjecture.

The outline of the tumulus, as shown in Fig. 10, is, so far as the broken lines extend, a restoration. It has been suggested that the analogy of other mounds of the class demands that the north-west and south-west corners should be rounded off. It is true that in most of the plans, the general outline is curvilinear and the tail' is sharply rounded. At St. Nicholas, the general outline is rectilinear, and the rounding of the corners would not materially affect this. To be consistent, the suggestion should include the north-east and south-east corners but at the former of these the north and east revetments so nearly meet that the rounding of the interval would scarcely be noticed on our plan. In a rectangular structure the corners are parts which soonest succumb to the buffettings of time, and this, in the case of the north-west and southwest corners at St. Nicholas, would be greatly accelerated, as the ground at each slopes away from the mound. It is questionable, however, whether the shape of the mound at St. Nicholas is exceptional. The west revetment at Uley is straight, and the horns at Stony Littleton are angulated¹-two characteristics of St. Nicholas. Aubrey's plan of the Wayland Smithy [Map]² and Dr. Thurnam's of the West Kennet [Map] remains³ resemble that of St. Nicholas, but are more elongated. Again, it is doubtful whether many of the published plans of the chambered tumuli are really correct. The only safe guide for the determination of the original form is the revetment, and too often this has only been opened out in a few places.

Note 1. Arch. and Nat. Hist. Soc. viii, p. 46. 2 Wilts

Note 2. Arch. Mag. vii, p. 315.

Note 3. Archaeo. xxxii, p. 405. VOL. XV.

It would have added greatly to the interest of the site if the revetment could have been kept open to view, but it was, as a rule, in too ruinous a condition to have stood the exposure. A short and very perfect length on the north side, and the whole of what remained on east or front, have, however, been kept open. The rest has been carefully covered up, but in such a manner that its line, as well as the form of the mound, are apparent at a glance. The face, as already stated, was generally leaning forward, and the removal of the talus of debris in front left it unsupported. To provide the necessary support, large stones were piled against it to form a steep embankment, and where the revetment was much reduced this was raised a foot or more above it. The embankment was then covered with earth, and more earth was stacked on the mound itself, most of the stones and the earth required for these being obtained from the talus. Finally, a course of large tabular stones, laid flat, was placed on the summit of the embankment to indicate the line of the buried revetment. About 16 ft. of the north revetment, east of the exposed length, was so completely pushed off its bearings that it had to be reconstructed, as also about 14 ft. at the west end of the south revetment-these are indicated by copings of slanting stones. The missing corners at the west end were made good by embankment.

At the opposite end of the mound, the revetment on each side of the approach remained to a considerable height, but much of that on the north was leaning forward in a dangerous manner. This was reconstructed, and to ensure the stones being replaced in their proper positions, each was previously numbered in chalk. In an easterly direction, these revetments were greatly reduced, and along the end of the north horn only one or two courses remained; while along most of that of the opposite horn, the field-ditch had obliterated every trace. It was necessary to restore these revetments to the existing height of the mound, and the new work is distinguished from the old by 'herring-bone' masonry, as shown in Figs. 6 and 8. The line of the new work along the end of the south horn is, of course, conjectural, and I am inclined to think that it is some inches in advance of the old. Within the chamber, the gap in the north-west corner between the two great side slabs has been filled in with dry walling, as indicated in Figs. 12 and 13. This was necessary to prevent stones and earth falling into the chamber but it may also be regarded as restorative work, for it was customary to fill in the gaps in these great structures in this manner. A small original patch is to be seen at the opposite end of the north slab. It is much to be desired that the chamber should be enclosed with an unclimbable iron fence, and this, I believe, Miss Cory contemplates having done.

The human remains and the small finds, consisting mostly of broken pottery and flint flakes, will be considered in a future paper, as, in the case of the first, it is not possible to do this until they have been carefully sorted and the skulls have been reconstructed as far as the fragments admit. Four of the skulls are now sufficiently advanced to show that they are of marked dolichocephalic type. (To be continued.)