of the bell-buoy clanging in my ears, I cleaned the sparking plug again, and again we got under way. The cliffs of Straight Point, with its ledges and rock-bound shore, were like a harbour to us. Over the dreary sandbanks of the Exe the bell was still tolling; is tolling now; but we had a sense of escape from it, and already felt ourselves as good as home. Once or twice more the engine stopped. It was not its fault: I was driving it badly. Darkness prevented me from tinkering about with it; but I found that by putting the propeller at neutral and letting it race for a few seconds the sparking plug would clear itself.

Not that the great argument came to an end therewith. The building and home-coming of the motor-boat were only two steps in the great argument, which still continues, and will. Cheated of breakdowns, which so far have not occurred, and bound to admit seaworthiness, the disciples of the Old Times complained loudly of the noise from the exhaust. Therefore, being otherwise well pleased, we brazened it out, and called the boat the Puffin. And then we stopped the noise.

THAT the site of one of the most famous battles ever fought on English soil should have been for more than a thousand years a mere matter of conjecture, is due not so much to the insufficiency of evidence as to the unscientific treatment of it at the hands of successive historians. The battle of Edington or Ethandune, which was the supreme climax to that famous struggle to the death between King Alfred and Guthrum, between Saxons and Danes, and between Christianity and Paganism, has been variously located in ancient Wessex, the favourite place so far being Edington, near Westbury, in Wiltshire. But the site is also claimed for Eddington, a Tything in the Parish and Union of Hungerford, in the county of Berks, and also for Edington, a small village on the Polden ridge above the tidal Parret in Somerset. In the absence of any direct statement as to precise locality either in the Anglo-Saxon Chronicle or in Asser, the historian and biographer of King Alfred, writing as a contemporary, it may be worth while to assay the testimony again, and with the aid of recent discoveries of place-names follow out the best inferential theory.

To begin with, we must remember that it is to old William Camden (1551-1623) that we owe the current report that the battle was fought

in Wiltshire, and it is from him that subsequent writers have copied with a fidelity that does more credit to their blind faith in what the Elizabethan antiquary wrote than to their cool judgment, or to any really serious conception of strategy. For Camden, learned as he was in many ways, was wofully inaccurate at times, especially when he drew a bow at a venture in philology or the derivation of place-names. For instance, we can hardly accept from him that "Summer" has ever been at the root of Somerset, the land of the "Sumorsaetas": or that Bristol or "Brightstow" was, to begin with, the "Bright" or famous town: or that the Cangi, a race of hypothetical giants living in Somerset, gave their name to such localities as Keynsham, near Bristol, or to Cannington, a village and hundred at the mouth of the Parret.

At the same time, there is a difficulty with modern and upto-date philologists, such as Mr Stevenson, the recent editor of Asser's works. His contention is that as the battle of "Ethandune" is so spelt in the A.S. Chronicle, and a similar spelling to describe the same place is found in Wiltshire documents in post-Domesday times, therefore the point is absolutely settled. He pretends to find in Anglo-Saxon, Norman, and English documents certain philological laws, like those of the Medes and Persians, which

alter not. He rejects the Somerset Edington because he finds that its Domesday equivalent (A.D. 1086) is Eduuinetune or Edwinetona. Now, if with Mr Stevenson we pin our faith on Domesday spelling, we shall find that the Wiltshire Ethandune is spelt Edendone. Further, in very early Glastonbury documents the Somerset Edington appears (A.D. 11001200) as Edindon also. The battle itself, in the twelfth century, may always have been spoken of and spelt as the battle of Edendone Edington, not Ethandune at all, as Henry of Huntingdon (A.D. 1150-1200) gives Edendune as the site, and Geoffrey Gaimer (c. A.D. 1150) gives Edenesdone in his 'Estoire des Engles.' But can we offer any explanation of the fact that the Domesday spelling of the Somerset place was Eduuinetona or Edwinetona? The history of the Glastonbury manor of Edington is easily traceable in the Rentalia and Costumaria of the great abbey, but it is never once spelt in the Domesday fashion. The Domesday spelling of the Somerset manor is a Domesday solecism, and some explanation of this may be found in the fact that Glastonbury had two manors-Edindone or Edington, so spelt early in the twelfth century, and lying on the Poldens, where it is well known; and another manor called Edwyne(s)ton, in Berkshire. The confusion may easily have arisen. Spelling in old documents is proverbially puzzling, as Mr Steven

son must know. There are nearly twenty ways of spelling Athelney, King Alfred's refuge. Combwitch, a modern placename, which Mr Stevenson once pronounced, in a letter to 'The Athenæum,' as an absolutely impossible variation of Cynuit, according to philological laws laid down by himself, is actually spelt in old documents as Cunyz, and the difference between Cynuit (where, it must be remembered, the great skirmish took place just before the battle of Edington) and Cunyz is not great. It is most unfortunate that, amongst able critics of ancient texts and documents, the question of the real site of the battle of Edington has been narrowed down to rather a pedantic assertion of a single philological theory. Too little attention has been paid to the features of the ancient geography at that date, which, with endless forests, meres, and floods, were so different from what they are now. Too little, indeed, has been given to the requirements of strategy and to the gradual evolution of a determined Danish plan of campaign, which had in view the absolute subjection of all Wessex, by means of land and sea co-operation. We feel sure, again, that if Camden had ever studied even a primitive map of the country (to say nothing of a modern Ordnance Survey), he would have hesitated about placing the site of the final battle of Edington forty or fifty miles from Athelney. But if Camden and his copyists are wrong, is it possible at this

belated period to turn to any new sources of information, and so to correct an old fallacy that has been engrained in our annals? The recent publication of old records, both in the Public Record Office and also in the 'Proceedings of Somerset Archæological Society,' has done much to facilitate modern enquiries, now more searching and exacting than they used to be. The task of tracing old manors and place-names, of following old roads and trackways, and, generally, of identifying the oldest features of our land, is far easier now for the historian than it was in Camden's time. For instance, what a light does the publication of the Athelney and Muchelney Chartularies shed upon the actual physical features of the refuge of King Alfred himself! Very often the charters will give us almost a cotemporary description of the original "herepaths" or strategic roads. An Athelney charter has certainly given us the Saxon Toteyate-i.e., the outlook island of King Alfred, which is "Borough mump, near Athelney; and a Wells charter has given us "Alfred's Road" close by, working in locally with the herepaths.

In the case of Asser, King Alfred's historian, the doubts cast upon his credibility by too destructive critics, who referred his works to a late copyist, have been dispelled. Asser's local descriptions have been found to chime in with facts, and a Wells charter which gave Bishops Lydeard to the bishop proved to be genuine enough.

VOL. CLXXXVIII.—NO. MCXL.

In 'The Fortnightly' of September 1899 the writer dealt with the question of the authenticity of Asser, arguing much from internal evidence afforded by Asser's own writings. Now that Asser has been seriously edited by Mr W. N. Stevenson, and his conclusions adopted by Professor Oman, it is hardly likely that his credibility will be challenged again. As stated in 'The Fortnightly' article on King Alfred's Country (1899), Somerset bears witness to Asser and Asser to Somerset as it was then physically and geographically, and old charters and wills (not excepting King Alfred's own will) have all given corroborative evidence.

To come to the chief points of this campaign, which, in spite of its remoteness, must be interesting, not only to the antiquary and delver amongst old MSS., but also to the modern strategist and "territorial" searching for precedents, it is clear that King Alfred was lord over Wessex, and was able to rule with a firm hand his domain from Winchester to Exeter. The Danes, whose ravages in the most eastern and accessible portions of England had elicited that piteous wail in the Anglo-Saxon Litany, "A furore Normannorum eripe nos, Domine," and, as Sir Henry Spelman reminds us, had caused the cessation of all Church synods, had begun to find their way westwards round the Cornish coast and so up the Severn Sea. In A.D. 845 a party of them had

2 L

tried to force the mouth of the Parret, a tidal river of the utmost strategic value to them, as it gave access to Langport and Somerton and the line of the Roman Foss way. But they were repulsed by Eahlstan, the warlike Bishop of Sherborne, Osric, the ealdorman of Dorset, and Eanwulf, ealdorman of Somerset.

The most real and the most dangerous attack upon the Parret mouth belongs to the years 877 and 878. It is not necessary here to dwell upon the events which led up to this culminating point. The strange thing is that King Alfred had baffled in 876 a descent of the Danes upon Poole harbour and Wareham, and in August 877, according to Sir John Spelman, had got the Danes by means of his superior skill or policy to evacuate Exeter, which they had occupied by a sudden coup de main after leaving Wareham. Up to this point his sway was unquestioned. Then, without any precise hints in the chroniclers of actual failure, the once powerful king becomes a fugitive in his own realm, living in a precarious fashion upon what he could get either from friend or foe in the marshes of North Petherton. From being head of a powerful host he becomes practically a guerilla chief, or as Sir John Spelman expresses it, "a flying Marius in the marshes of Minturnæ."

To account for this extraordinary collapse we must look for sufficient causes, as these are not directly emphasised in

the Anglo-Saxon Chronicle, and it is enough to point out here that King Alfred's kingdom of Wessex had been threatened in the interval not simply along the south coasts of Dorset and Devon, as before, but by way of the Severn Sea. We hear of Hubba, the great Danish chief, passing his winter in Demetia and South Wales, where sufficiently reliable tradition says that he made his quarters in or near the ancient Forest of Dene, "so called from the Danes in King Alfred's time," as Giraldus Cambrensis tells us. There was plenty of booty to be had in the rich Monmouth monasteries, which seem to owe their desolation to this period, and there was abundance of iron to be found in the Forest of Dene, where formerly the Romans had worked; and, no doubt, the weapons and armour of the Danes had to be replenished.

Moreover, it is quite clear that for a piratical host operating as that of the Danes did, making sudden swoops and unexpected forays, the Severn Sea was most admirably suited as a base. Even in the winter months the "trajectus was comparatively easy, and Hubba was able to keep the whole coast of North Somerset and North Devon in continual alarm. The strategic value of the occupation of the Usk and Caerleon is more and more apparent. The kingdom of Wessex was being assailed in its most vulnerable quarterie., along the Severn Sea; and