rather than the reverse. Thus the marginal entry in Matthew Paris and the Breton tradition are, though evidence of doubtful value, both independent of the documents of 1216, and find a parallel in the chronicle of Coggeshall, whose importance is indisputable.

2. The story of Arthur's death which is most likely to be true, and is corroborated by other evidence, is contained in the annals of Margam. The condemnation of John is an integral part of this story, which has no connection with the documents of 1216 and is probably due to William of Briouze.

3. Too much stress has been laid upon the argument from silence.

I.

Within thirty or forty years of his death that great southerner Richard the Lion Heart had become a peculiarly English hero of English romance-romance full of confused reminiscences and picturesque nonsense, which in its amplified anti-French form was used by Shakespeare; and the notorious John suffered by comparison in popular history. Most of the popular version of John's misdeeds may be put on one side; but the more critical narrative of Holinshed is a suggestive starting-point for a study of the medieval tradition. Holinshed gives his authorities. The story of Arthur's interview with Hubert is based on a contemporary Essex chronicle of Coggeshall. Holinshed repeats the three or four rumours made current by Matthew Paris in his Historia Anglorum,2 that Arthur died of grief, or was drowned in trying to escape from the tower of Rouen, or was killed by his uncle. The most popular version of Arthur's death is unknown to Holinshed,

1. See G. Paris in Romania, xxvi, 357, 387. Compare Bishop Bale's long since forgotten play about King John, which Shakespeare is said to have used.

2. Ed. Madden, ii, 95.

and therefore to Shakespeare. Hence in the famous play, the Hubert scene naturally becomes the central theme.

There was a Breton tradition also, which was familiar in the fifteenth century and was worked into the narrative of the learned Breton historians of the seventeenth century. 1 According to this version the barons and bishops of Brittany assembled in great numbers and charged John with the murder fifteen days after it was committed. On the strength of this charge King Philip of France condemned the English king to lose all his possessions. So far as this story is true, it can be traced, as M. Bémont pointed out, to the events described by the Essex chronicler, Ralph, abbot of Coggeshall, to whom I have referred.

King John captured Arthur at the castle of Mirebeau on 1 August 1202. Arthur was between fifteen and sixteen years of age-nearly a man in those days-and had been invested by Philip with all the Angevin lands outside Normandy. At the time of his capture he was besieging his grandmother. He was taken to Falaise and imprisoned in the tower. John is said to have promised that if, with the aid of William des Roches, the most powerful baron and official in Maine and Anjou, he succeeded in defeating Arthur, he would act on William's advice. His trickery after the successful march on Mirebeau and his cruelty to the prisoners cost him the allegiance of William and of the barons of the west. They joined with the Bretons and the rebels of Poitou. Some of the Normans were won over. 2

The abbot of Coggeshall is the sole authority for what happened at Falaise. John's counsellors saw that so long as Arthur was kept in Falaise, away from his followers, yet

1. See Bémont, Revue Historique, xxxii (1886), 290-300; Stubbs, Introduction to Walter of Coventry, ii, p. xxxii. Several continental chroniclers refer to the rumour of Arthur's murder, but their evidence throws no light on the facts.

2. Vie de Guillaume le Maréchal, iii, 167-170; Coggeshall, p. 139. Above, p. 232.

3. Ibid, pp. 139–141.

safe and well, John was in danger. It must be remembered that the king was already under sentence of deprivation by the French court, on account of the appeal of the Poitevin barons. If the alliance was not to be overwhelming Arthur ought either to be handed over to William des Roches or to be put out of the way. Some of John's friends suggested mutilation. In his anger at failure, after the only brilliant military achievement of his life, John agreed, and sent two servants to Falaise, where, his feet fettered by a triple chain, the young man was guarded by Hubert de Burgh, the chamberlain. Hubert, moved partly by the agony of Arthur, partly by the folly of the deed, prevented John's agents from accomplishing the royal command. Yet he felt also that the only way to coerce the Bretons was to convince them of Arthur's death. What folly there might be in mutilation or murder lay in the fact that John's subjects, especially his knights, would refuse to serve a parricide. Hubert announced that Arthur had died. For fifteen days (we see here the fifteen days of the Bretor story) the rumour spread. The place of Arthur's burial was known also. Then the Bretons, fully roused, swore that they would never cease their attacks on the king of England after this atrocious deed. They believed that Arthur had been murdered. It is not at all unlikely that they held a solemn assembly; the Coggeshall narrative rather implies common action. In this case the chief facts of the Breton version would be true, and the fifteenthcentury and later writers were following veracious but obviously independent annals in their detailed account of the gathering at Vannes. The error simply lay in this, that Arthur was not yet dead.

This explanation is the more probable because from that time Arthur disappeared. Hubert, when the danger increased rather than diminished, announced that he was alive, but the Bretons could have no proof of this. They would naturally prefer to believe that Arthur was dead, if he was not handed over. Philip and they clamoured for

his release and offered hostages in vain. Their scepticism is expressed distinctly in the charter of King Philip in which he refers to Arthur if he still lives.' Till the spring of 1204 this scepticism was maintained; then it became certainty that Arthur was dead; but there was no proof. The semi-official chronicler Rigord of St. Denis, who lived till about 1206, makes no mention of it. A few chroniclers tell us that Arthur was removed to Rouen; and no doubt, as time went on, this fact became common knowledge. But after that all was darkness and vague rumour. Only here and there-e.g. by the chronicler of Tours2was Arthur supposed to have been killed. In 1204 Philip refused peace, partly because he was confident of success in war, partly and especially because he had heard that Arthur had been drowned in the Seine. Many years later even Matthew Paris, who was not exactly friendly to John, can only give the various stories of his death and hope doubtfully that the story of murder is not true. Gradually, in popular talk Arthur's fate became subject to the variations of time and place and incident which control all mysteries.

Such was the main historical tradition concerning the relations between John and his nephew. Putting aside other evidence as valueless, M. Bémont has urged that it is sufficient to disprove the story that John was condemned, a second time, for the death of Arthur. It certainly does not prove it, but it is hard to see how it can be said to do more. The condemnation of John ought to be considered together with the question, When did Philip become morally certain of Arthur's death by murder? The orthodox view is as follows: John must have been

1. Delisle, Catalogue des Actes de Philippe-Augustus, no. 783, p. 177, October, 1203, before Château-Gaillard, charter for Guy of Thouars; Bémont, Revue Hist., xxxii, 42.

2. Historiens de France, xviii, 295.

3. Saeviebat autem permaxime pro nece Arturi, quem in Sequana submersum fuisse audierat: Coggeshall, p. 145.

2

condemned, if at all, in 1203; and, as Philip was uncertain of Arthur's fate in April 1204, John could not have been condemned at all. Now the only serious reason for the statement that John must have been condemned, if at all, in 1203 is that Philip continued the war in 1203, and sentence must come before the punishment. This in its turn seems to imply that Philip would not have invaded Normandy in 1203, if John had not been condemned. It is true that the later writers, looking back, are so much impressed by the crime that they say it caused the loss of Normandy, as indeed it did to a large extent. Philip was urged on by indignation. One or two very important witnesses, as we shall see, imply that Normandy was escheated because of the sentence. Indeed, if sentence was passed, this must have been true also. But all these considerations are irrelevant to the fact that Philip, while still uncertain or ignorant of Arthur's fate, invaded Normandy in 1203, and would have done so in any case. The evidence for the condemnation is not invalidated because some of the witnesses thought that it caused a war already in progress. The truth is that Philip and John were at war and that there was no break. It is certain that Philip regarded Normandy as escheated in 1202, together with Poitou and the other possessions of King John. There is no hint that the military operations from the opening of war in 1202 to the surrender of Rouen in June 1204 were not regarded as continuous. Rigord says

3

1. Revue Historique, xxxii, 55.

2. The anonymous chronicler of Laon, who is especially interested in Anglo-Norman history, puts the case exactly from the retrospective standpoint: 1203, Iohannes rex Anglie Arturum . . . crudelissime iugulavit. . . . Guera inter regem Francie et regem Anglie fit solito gravior (ed. Cartellieri, p. 61).

3. This is proved by the papal letters of 1203 (Potthast, no. 2013) taken with Arthur's letters of July, 1202 (Layettes, i, 236, no. 647).