But perhaps the most convincing negative evidence is supplied by Gildas, to whom I have already referred. This historian, a fellow-countryman and contemporary of Arthur, was either ignorant of his existence or thought him not worth mentioning. Now Gildas, as we are told by William of Malmesbury, 'took up his abode' at Glas...o...b..ry 'for a series of years.' If Arthur died, as was supposed, in the year 542, and Gildas was born in 520, the historian must have been twenty-two years old when the king was buried under the description of 'the famous King Arthur,' _inclytus Rex Arthurius_. Gildas might have been present had this taken place as represented, or at any rate must have heard from his friends the monks of what could not fail to be of interest to the British historian. But neither Arthur's death nor his life appealed to Gildas. Thus we must discredit both the Camel and Glas...o...b..ry as connected with Arthur's death and burial.

IV

TOPOGRAPHICAL a.s.sOCIATIONS

I do not propose to follow in detail the romancers of the twelfth and succeeding centuries, excepting where they may be taken in concurrence with surviving structures and geographical peculiarities. I have said something in this sense both of the Cornish and the Scottish localisation of Camlan. Turning from the conclusion of Arthur's career to the beginning of it, I must again have recourse to Geoffrey of Monmouth, a writer who sometimes finds the corroboration which he always needs.

Uther Pendragon, King of Britain, held a festival in London, or, according to another account, at Winchester, at which were present Gorlois, King of Cornwall, and his wife Igerna, 'the greatest beauty in all Britain.' Uther was more attentive to the lady than was approved by her husband, who abruptly left the Court and returned to Cornwall, taking his wife with him. Uther followed. Gorlois deposited Igerna in Tintagel, 'upon the sea sh.o.r.e, which he looked upon as a place of great safety. But he himself entered the castle of Damelioc to prevent their both being involved in the same danger if any should happen.' Damelioc is described as a strong 'castle,' having many issues out. According to the legend, Gorlois hoped here to receive succour from Ireland. In this place Gorlois was besieged by the superior forces of Uther, and was slain fighting outside its ramparts. Igerna was apparently secure in Tintagel. 'For it is situated upon the sea and on every side surrounded by it; and there is but one entrance into it, and that through a straight rock, which three men shall be able to defend against the whole power of the kingdom.' But though Tintagel was impregnable, the lady was not. Uther obtained admittance into the castle while Gorlois was in Damelioc, and Arthur was the result.

According to tradition, Uther was transformed into the likeness of Gorlois by the arts of Merlin, and the King of Britain admitted under a misunderstanding to the domestic privileges of the King of Cornwall. The sceptical may hesitate to accept this explanation of the error of Igerna, but no doubt it was furnished by the lady herself, who could scarcely fail to have been acquainted with the facts. My purpose in alluding to the story is rather local than personal. The description of Tintagel might serve at the present day. Part of the castle is on a lofty and precipitous peninsula commonly known as the Island, which has only a narrow connection with the mainland, which few could defend against many. It is obvious that it was on the Island that Igerna was placed and Arthur begotten.

The Castle of Tintagel is so closely connected with the Arthurian legend that a few particulars concerning it must be introduced. At Tintagel, according to tradition, the Kings or Dukes of Cornwall had their residence before the coming of Caesar. The place was formerly known as Dundagell, and is supposed to be indicated by the name Donecheniv, which is to be found in 'Domesday Book,' and according to Gilbert means the fort or castle with the chain. This is the earliest reference to Tintagel, if it be one, which I have been able to discover. The allusion to the chain is appropriate. There is evidence that the chasm which separates the insular part of the castle from that on the mainland was formerly crossed by a drawbridge. This is likely enough, for the chasm was evidently once narrower than at present, having been enlarged by the falling away of the cliff, while the buildings on the mainland and the island are opposite and near to each other, as if they had at one time been connected. Leland in his 'Itinerary' describes the castle as it existed in the time of Henry VIII. In the earlier part of his work he refers to a drawbridge as connecting the two portions of the fortress; in a later part he states that the island could be reached only by long elm trees laid for a bridge.

Other writers refer to the bridge. Carew in his 'Survey of Cornwall' in 1602 states that this was in existence one hundred years before he wrote, and Norden, a writer of about the same date, says that it was there within living memory. It is obvious that the historical bridge belonged to the buildings parts of which still exist. The allusion in 'Domesday Book,' if correctly interpreted, must relate to an earlier structure, for there is reason to believe, as I shall presently show, that no part of what now remains existed at the time of the Conquest. Nevertheless, there was probably at that time some mechanism with a chain which gave access to the island from the adjacent cliff.

It is not my purpose to give a detailed description of Tintagel Castle, such as may be found in many works relating to the locality;[20] but a few words bearing upon the question of its hypothetical a.s.sociation with Arthur seem called for.

The site of the castle is remarkable: it is partly on the mainland and partly on a peninsula which from time immemorial has been known as the Island. This is separated from the mainland by a deep chasm which is evidently in process of enlargement, or, in other words, was once narrower than it is now. The island, which is bounded by lofty precipices, is connected with the mainland only by a narrow ridge, which rises steeply from the sea, traverses the chasm, and gives access to the island by a narrow path cut in the face of the cliff, which now, as in ancient days, might be defended by a few against many.

To take first the insular part of the castle, which no doubt was the original place of retreat and defence, the site may be a.s.sociated with that of many prehistoric fortifications of earth or stone, the remains of which are to be found on the Cornish coast. The ancient engineers habitually selected a precipitous peninsula, inaccessible from the sea, with a narrow neck, across which they made barriers to protect against attack from the landward side. Thus Tintagel Head was selected as a place of defence, if not by prehistoric engineers, certainly in accordance with prehistoric methods. The buildings at present on the island are less extensive than those on the mainland. There is no evidence that any part of them is anterior to the Plantagenets. An arch which forms the gateway of the outer wall is distinctly though bluntly pointed, and must be later than the Norman period. A bluntly pointed arch, known as the Iron Gate, is also to be seen in a wall which protects what was apparently once a landing place. Outside the enclosure of the castle are the wind-swept remains of a little chapel which should be that in which Merlin vainly sought repose. Old it undoubtedly is, but the most credulous could scarcely attribute it to the sixth century. In construction it resembles the rest of the insular part of the castle, being not too solidly built of roughly quarried unsquared slates. There is nothing of architectural style to determine the date, but the walls resemble the others and may be presumed to be like them of the early Plantagenet time.[21]

The buildings on the mainland give more scope for discussion. These are placed on a high, narrow elevation which rises out of a gorge: this elevation, which is steep on one side and precipitous on the other, rises above the level of the buildings on the island, with which at one point they may easily have communicated. The structures on the mainland consist of two walled enclosures on different levels, connected by steps. The lower and larger is supposed to be the courtyard, the higher the keep, and indeed they do not admit of any other interpretation. The courtyard presents towards the land the remnants of a great gateway, while towards the sea the wall has fallen, exposing a precipice where once the wall stood. The gateway is of especial interest: what remains of the arch is suggestive that it once was pointed, and I have the evidence of an intelligent mason who lives hard by, and who was familiar with its condition twenty years ago, that though then broken it retained enough of the curve to indicate that originally it was bluntly pointed, and resembled in construction those still to be found on the island. I may add that I have seen a drawing executed by Mr. Sturge, about sixteen years ago, from which the same inference is to be drawn. I may draw attention to a photographic reproduction of a print of the castle as it was about 300 years ago, when the gateway was complete (see fig. 1, p. 62). The arch in question appears to be less flat than it should be were it Norman, though the scale of the drawing is too small to display a distinct point. It has been supposed that the upper enclosure, known as the keep, is older than the lower or courtyard, and the late Prebendary Kinsman thought he had found traces of Roman methods in a projecting course of flat stones which traverses the upper part of one of the walls; but I am inclined to agree with my friend the mason, who considers the projection to be of English invention, designed to protect the wall from weather and give finish to its top.[22] The keep is connected with the courtyard by a flight of steps, as if the two formed part of the same design, while the masonry of the two portions is exactly of the same character, as if they were coeval.

That the upper and lower enclosures formed parts of the same design, is sufficiently evinced by the drawing which has been reproduced.

The insular part, though showing similar work and material, is in better preservation; indeed, it is not easy to doubt that it is considerably later, though belonging to the same architectural period. The pointed arches indicate that neither the continental nor the insular part were constructed before the introduction of this form of arch. The pointed arch gradually superseded the round arch during the reign of Henry II.--1154-1189--and did not become general until quite the end of this period.[23] There are indeed pointed arches in the church of St. Cross, near Winchester, which are supposed to date back as far as 1136, but this appears to have been a solitary instance some 50 years earlier than the general employment of the style. It may fairly be presumed that neither portion of the existing buildings dates back further than the twelfth century, while the insular portion is probably less ancient than that on the mainland. There is evidence that there were buildings on the island at an earlier date than can be ascribed to those now existing.

Geoffrey of Monmouth, who was made Bishop of St. Asaph in the year 1152, and probably wrote earlier, describes the place, in words which I have already quoted, as he supposes it to have been in the time of Uther Pendragon. He calls it 'the town of Tintagel, a place of great safety. For it is situated upon the sea and on every side surrounded by it; and there is but one entrance into it, and that through a straight rock, which three men shall be able to defend against the whole power of the kingdom.' Not to insist upon Uther, it is clear that Geoffrey intended to describe the place as it was before his own time, and, by unavoidable inference, before the buildings at present on the island were constructed. For it is almost certain that none of them existed in the time of Geoffrey--quite certain that none of them were built before his birth. It is to be noted that this writer makes no allusion to the part of the castle on the mainland, which, though probably older, was presumably not made when he wrote. We cannot but infer that before Geoffrey's time there was some sort of fortification on the island, which was replaced by the existing structure; and this inference is supported by the name under which this place is referred to in 'Domesday Book,' if Gilbert is correct in his interpretation of it as 'the fort with the chain.' The evidence that Tintagel Head was used as a stronghold before the present buildings were made lends credibility to the tradition which connects Arthur with Tintagel, though none of the present walls were constructed until at least 600 years after his death.

With the great gateway at one end and the exposed precipice at the other, the courtyard corresponds with a description written in the thirteenth century, and designed to present the state of the castle in the time of Arthur.[24] Through this gateway, according to the romance, rode Uther and Merlin, and within these walls Arthur was begotten. It is much to be regretted that the building so adapted to the story had no existence in the time to which it relates. It may be objected also that the romancer has made a capital error in placing the adventure on the mainland, and a minor error in a.s.signing the same position to the chapel. It is to be presumed that the story-tellers long subsequent to Arthur's time adapted the legends relating to Tintagel somewhat loosely to the building as it existed in their own.

The 25-inch Ordnance Map represents the continental part of the castle as built upon the site of a camp. It is with great diffidence that I venture to question this interpretation of a trench which runs parallel to, and close to, the south wall of the castle. This trench must, I think, be accepted as having been made simultaneously with, or subsequently to, the building, for it evidently bears relation to the great gate and to an otherwise unprotected wall, of which it formed an outer defence. Sir John Maclean calls this a moat. If this means no more than a defending ditch I am of his opinion, but if a moat should hold water the term is inapplicable, for the fosse is on such a slope that water never could have remained in it. As to the camp theory, it may be observed that there are undoubted remains of a camp within a quarter of a mile, close to the church, and it is unlikely that two camps would have been constructed in such proximity.

[Ill.u.s.tration: FIG. 1.--_Tintagel Castle as represented by Norden_, 1584-1600.]

[Ill.u.s.tration: FIG. 2.--_Tintagel Castle from 25-inch Ordnance Map_]

I insert a drawing (p. 62), to which I think much interest attaches. It represents the castle as it was about the year 1600--roughly speaking, 300 years ago. It is a copy made by photography of a print in 'Norden's Speculi Britanniae Pars,' a book now in the British Museum, formerly in the Royal Library. It is dedicated to James I., and appears to have been written at the end of the sixteenth century. The date 1584 has been doubtfully a.s.signed to it: we may safely refer it to the end of the sixteenth century. Norden was born in 1548 and died in 1626. He was Surveyor of Woods to James I., and evidently regarded architectural accuracy more than pictorial effect. The drawing shows the landward part as extending further seawards than at present, while it indicates a place where parts of the insular buildings had recently been engulfed. The great gateway on the mainland is entire; the keep and the lower court nearly so; while the relation of the three as parts of the same building, presumably built at the same time, is unmistakable. The buildings on the island are much as they are now.

I append what may also be of interest--a facsimile of as much of the 25-inch Ordnance Map as relates to Tintagel Castle. For permission to do so I have to thank the Director of Her Majesty's Stationery Office, to whom also I am indebted for similar permission touching the maps of Damelioc, Kelly Rounds, and Cardinham Castle, to which I shall presently draw attention.

So much and so little for the Castle of Tintagel and its relation to King Arthur, who certainly never saw an arch or a stone of the existing building, and could not have been begotten in a hall which was not made until many centuries after his death. What took place, and where, before the hall was built are other questions. This does not necessarily detach Arthur from Tintagel. It is probable that Tintagel Head was a place of strength and of retreat in prehistoric times, as were many headlands on the Cornish coast. No earthworks were needed to secure it, as it had been rendered impregnable by nature. The Celts in the sixth century, and in Cornwall, though they must have had skill in stone-carving if some of the existing crosses are correctly attributed to this period, were probably not castle builders, unless the term 'castle' be applied to earthworks, as is indeed still the custom. If Arthur was the lord of Tintagel, as is indicated by an immemorial tradition, which we may, without violence to probability, accept, nothing remains of this his dwelling-place excepting the immortal ramparts which will be for ever a.s.sociated with his name.

_Dameliock_ or _Dimilioc_ still exists as a formidable earthwork, and retains the name.[25] It is mentioned in 'Domesday Book' under the term Damelihoc. 'Domesday Book' is a mere rent-roll and does not deal in descriptions, but the name is significant if Gilbert[26] is correct in his interpretation of it as 'the place of battle.'

Damelioc Castle is a British camp of great strength. It lies about eight miles from Tintagel Castle by the modern roads, though in earlier time the distance may have been as much as ten miles, which is a.s.signed by tradition as the distance of Damelioc from Tintagel. It is in the parish of St. Kew, close to the road which connects Tintagel with the mouth of the Camel, and, taking a wide scope, may be said to lie between the north of Cornwall and the south of Ireland. The work is said to extend over an area of twelve acres. It once consisted of three concentric ramparts, of which two remain effective if not complete, while portions of a third and outer are still to be seen, though much of it has yielded to the invasion of agriculture. Outside each of the two nearly complete ramparts, of which the outer is the higher, lies a ditch. This rampart is from 20 to 30 feet high, measured from the bottom of the ditch;[27] it is about 11 yards thick and the ditch 10 yards wide. The inner circle is less perfect than the outer, or more properly the middle, little remaining except the ditch, which is about 5 feet deep. The enclosure, measuring from the inside of the middle rampart (the outermost is not complete enough to reckon by), has a diameter of about 170 yards and a circ.u.mference of about 530 yards.

The enclosure formed by this rampart would, according to Colonel Mead, whose a.s.sistance I had in examining the fortification, give comfortable accommodation to 2,000 men, supposing them to be besieged for a week or ten days, while under temporary pressure 5,000 people might be crowded into it. Were the outer circle restored according to the indications afforded by its remains, it is obvious that the camp would hold a much larger number than that referred to as capable of being contained within the middle defence. Like others of the Cornish earthworks, it stands on a commanding elevation among hills which are higher than itself.

It has been seen that Tintagel and Damelioc are exactly adapted to the story of which they are the scene. Either the story must have had some foundation in fact, or the inventor of it must have possessed extensive and accurate knowledge of the topographical features of this remote part of the British Isles.

[Ill.u.s.tration: FIG. 3.--_Reproduced from the 25-inch Ordnance Map._]

The Romancers of Britanny may easily have heard of a place so well known as Tintagel, and woven it into their fictions; but Damelioc seems to have attracted little attention, though mentioned by Gilbert in 1838, until it emerged from obscurity to find name and place in the last edition of the Ordnance map. Thus unknown or disregarded, it would scarcely have been selected as the scene of a purely imaginary romance. To me, the finding of Damelioc where and what it should be according to the story is an indication that this was dictated by something more substantial than imagination, though this faculty no doubt had much to do with its embellishment.

I have already quoted from the Welsh Triads a.s.signed to the sixth century a reference to Arthur as 'the chief lord at Kelliwic,' and have referred also to other Welsh compositions, probably of little less antiquity, in which Kelliwic or Celliwig is spoken of in the same connection. Professor Rhys finds in the Triads an account of a raid made by Mordred[28] upon Arthur's Court, apparently in Arthur's absence, where the intruder left neither food nor drink unconsumed so much as would support a fly, and where he outraged the Queen. This is said to have occurred at Kelliwic in Cornwall, though it must be admitted that the a.s.sociation of the northern king with the southern fortress is suggestive of doubt. Kelliwic is elsewhere referred to as a place from which a certain marksman of exceptional ability was able to hit a wren in Ireland. Dismissing this as one of the super-additions to which tradition is liable, I revert from the archer to the king. If there be any truth in the tradition which places Arthur's court or camp at Kelliwic, we ought to find some trace of it. If Kelliwic could be found as a place of defence in the Arthurian country, we might at least say that the coincidence was remarkable, unless the tradition had some substratum of fact. Now I venture to suggest that we have Kelliwic still with us under the name of that remarkable earthwork known as Kelly Rounds.

_Kelly Rounds_ or _Castle Killibury_ is about five miles from Damelioc, to which it bears a general resemblance, though possessing only two ramparts, with no traces of a third. The work is situated near the road between Camelford and Wadebridge, about 2-1/2 miles from the latter, which is a well protected port. It consists--or rather I should say consisted--of two concentric circles, each with rampart and ditch. It is obviously a British camp. A road now cuts it into two nearly equal parts, of which that on the south has been nearly obliterated, while the northern segment is comparatively uninjured. The ramparts, of which the inner is the higher, present a maximum height of perhaps 15 feet, judging roughly by the eye.

The diameter of the remaining semicircle is about 210 yards, measuring from the inside of the outer rampart, while the semi-circ.u.mference in the same position is 290 yards. On the west side are the traces of an outwork, or partial enclosure, which was evidently designed to protect the entrance.

The extravagance of the archer who 'shot with a l.u.s.ty longbow' from Kelliwic to Ireland is not quite without significance, for it may be held to show that Kelliwic, like Kelly Rounds, was opposite the Irish coast.

[Ill.u.s.tration: FIG. 4.--_From the 25-inch Ordnance Map._]

We may with some confidence identify Kelly Rounds, or Castle Killibury, with Kelliwic, and discern in it, as in Damelioc, a definite a.s.sociation with Arthur.

A place to which the name of _Caradigan_ is given is prominent in Arthurian lore. This has been interpreted as Cardigan, the ancient designation of Cardiganshire being Keridigion.[29] But Mr. E. G. B.

Phillimore, who is a great authority on ancient Welsh literature, considers that Caradigan is not Cardigan, but Cardinam, now known as Cardinham, a considerable, though much damaged, earthwork near Bodmin. In this interpretation Mr. Phillimore apparently has the approval of Professor Rhys. If Caradigan is Cardinham, this was one of the places where Arthur held his Court. It was at Caradigan that Enid was wedded to Eric by the Archbishop of Canterbury, in the presence of Queen Guenevere.

It was to Arthur's Court at Caradigan that Lancelot brought his newly-married wife, Iblis.

The doings at Caradigan are obviously mingled with fiction, if not wholly fict.i.tious. The Archbishop of Canterbury was not yet, and Eric as a knight of mediaeval chivalry is, like the Archbishop, an anachronism; but there is something in a name, and Caradigan a.s.sociates Arthur with the Cornish stronghold.

Cardinham Castle, as it is called, though far inferior in size and distinctness to Killibury and Damelioc, is worth more notice than it has yet received. About five miles from Bodmin, on the edge of Cardinham Moor, lies Old Cardinham, now represented by a solitary farm-house. In a field behind the house stands an earthwork, of small extent but great natural advantage. It is situated, like Damelioc and Kelly Rounds, on high ground among hills which are higher than itself, but not near enough to command it without artillery. This stronghold or place of defence displays the remains of one rampart enclosing an ovoid or irregularly elongated s.p.a.ce on the side of a hill, within which the experts of the Ordnance Survey discern a small inner circ.u.mvallation. In designing the enclosure the natural slope has been made use of to co-operate with the rampart on the north side, while the rampart on the south is wholly artificial, much broken, and in places obliterated.

[Ill.u.s.tration: FIG. 5.--_From the 25-inch Ordnance Map._]

The partial destruction of the south wall makes the enclosure incomplete, and gives it a horse-shoe shape. The entire circuit along the tops of the existing and nearly obsolete ramparts is about 267 yards, and this comparatively small circ.u.mference encloses a narrow and elongated s.p.a.ce of relatively small capacity. The surface is irregular, and may once have had buildings upon it, of which there are now no remnants. This small but well-protected enclosure seems to have been better fitted for a fortified residence than a resort for an army. It may conceivably have held the residential quarters of a Cornish chieftain in the sixth century, and its legendary a.s.sociation with Arthur may not impossibly have had some foundation in fact.

V

CONCLUSIONS